Sagility Limited — Call Transcript 2026

Apr 2, 2026

==> picture [134 x 32] intentionally omitted <==

Date: April 02, 2026

To,

The Manager Listing Department National Stock Exchange of India Limited (NSE) Exchange Plaza, 5th Floor Plot No. C/1, G-Block Bandra-Kurla Complex Bandra (E), Mumbai - 400 051 Symbol: SAGILITY

The Manager Listing Department BSE Limited (BSE) Phiroze Jeejeebhoy Towers Dalal Street Mumbai - 400 001 Scrip Code:544282

Dear Sir/Ma’am,

Subject: Transcript of Investor & Analyst Day held on March 25, 2026

In continuation of our earlier communications dated March 13, 2026 and March 25, 2026, please find attached the transcript of the Investors’ and Analysts’ Day held on Wednesday, March 25, 2026.

• This information is also available on the Company’s website https://sagility.com/investor relations/financial-summary/

This is for your kind information and record.

Thanking You,

For Sagility Limited

SATISHKUMAR Digitally signed by SATISHKUMAR SAKHARAYAPATTAN SAKHARAYAPATTANA A SEETHARAMAIAH SEETHARAMAIAH Date: 2026.04.02 16:35:33 +05'30'

Satishkumar Sakharayapattana Seetharamaiah Company Secretary & Compliance Officer M. No. A16008

Encl:a/a

Sagility Limited

(Formerly Sagility India Limited, earlier Sagility India Private Limited)

Registered Office - No. 23 & 24, AMR Tech Park, Building 2A, First Floor Hongasandara Village, Off Hosur Road, Bommanahalli, Bengaluru – 560068, Karnataka, India Corporate Identification Number: L72900KA2021PLC150054

Tel. No.: 080-71251500, E-mail : investorservices@sagility.com, Website: www.sagility.com

==> picture [146 x 39] intentionally omitted <==

Sagility Ltd.

Investor & Analyst Day Transcript Wednesday, 25[th] March 2026

MANAGEMENT

• Ramesh Gopalan - Group CEO and Managing Director

• Srinivas Mattapalli – Executive Vice President, Group Chief Financial Officer

• Abhishek Kayan – Deputy Chief Financial Officer

• Roopam Narayan - Executive Vice President, Solutions & Practice

• Srikanth Lakshminarayanan - Senior Vice President, Healthcare Practice

• Krithika Srivats - Senior Vice President, Clinical Practice

• Madan Moudgal - Executive Vice President, Chief Digital Officer

• Ram Mohan Natarajan - Senior Vice President, Business Transformation

• Chris Shiffert - Chief Growth Officer

• Dhaivat Mehta – Vice President, Global Brand and Corporate Communications

QUESTIONERS

• Atul Mehra – Motilal Oswal Financial Services

• Pallavi Deshpande – Sameeksha Capital

• Shreesti Rastogi – Canera Robeco AMC

• Siddharth Vora – HSBC Mutual Fund

• Anil Nahata – Parami Financial

• Chetan Shah – Jeet Capital

• Namit Arora – INDGROWTH Capital

• Pallavi Deshpande – Sameeksha Capital

• Ravi Menon – Macquarie Capital

• Varun Gandhi – Finavenue Growth Fund

Page | 1

Dhaivat Mehta:

Good afternoon, this is Dhaivat Mehta, Vice President of Global Brand and Corporate Communications here at Sagility. And on behalf of everyone at Sagility, it is our pleasure and privilege to have you amongst us here today. I hope you've been able to grab a bite to eat. We really appreciate you being amongst us. I know all of you have busy days, and so we really do appreciate it. Well, today, we have several members of our leadership team here, and we're going to walk you through some of our strategic roadmap and how we intend on continuing to create long-term sustainable value.

We are going to have a Q&A session. Team, if you can go to the next slide, please. I want to make sure that we share with our delegates here. So, this is a QR code, and there are several QR code stands across the room as well. If you have any question that comes up as we go through the presentations, please feel free to put in your question on this QR code link. We will have a dedicated Q&A session right at the end.

We will take all those questions that you have put forward, but also, we will have a team that will go around with mics, and so you will be able to ask questions that way as well. Of course, if you go to the next slide, it's important that we take a moment to look at the safe harbour statement regarding forward-looking statements. So, I will pause for three or five seconds and we can look at the statement together. Okay. With that, I won't be taking too much of our time.

Let me hand it over to Ramesh, our MD and CEO, who will take the next session. And we have multiple sessions thereafter. We have a tea break. We have a long day here today. We are going to be here together till dinner and cocktails. So, look forward to knowing many of you better.

Ramesh, over to you.

Ramesh Gopalan:

Okay, thanks. Thanks Dhaivat. Good afternoon. Welcome everyone. Lots of familiar faces. So like Dhaivat said, it's a long day. We are going to continue all the way until 7. Let me quickly tell you what's planned. So, this is the broad structure of what we intend to present.

So, as you can see we will start with my session, but the whole idea here is I have met almost all of you. You have heard from me several times, but today is the opportunity for you to hear from my team. So we will dive deeper into some of the topics. I want you to walk away with a good understanding of our capabilities, our go forward strategy and how we plan to execute on that strategy over the coming years.

Page | 2

We have a tea break between 4 and 4.30 but a lot of our presentations include demos. So hopefully you will get a better feel for some of the products and solutions that we have built and like Dhaivat said at the end of the day we will have a 45-minute Q&A session. So please post your questions through the QR code and also, we will have mics around at the end for further questions. Again, just a placeholder for those of you who don't know us. Many of you have seen this slide before. We are a leader in healthcare operations transformation.

So that is what we do. All of our clients are U.S. healthcare clients. Two major segments Payers and Providers. Payers constitute about 87%, 88% of our revenues. Providers about the balance 12%, 13% and we go to market on three pillars. One is a deep domain expert experience and throughout this afternoon that is the key theme that you will keep hearing from the rest of my team. You will get a first-hand experience of how deep in domain we are as experts in U.S. healthcare and the second thing you will also get to see the breadth and depth of a service portfolio. So, I've told many of you in the past that when it comes to especially the payer value chain, we touch almost all the functions in the payer’s ecosystem.

So, and just not touch cursorily but we have deep expertise in each of those functions and that sets us up for some of the future work that we are doing and the third part again over the last at least 10 plus years we've been focused on tech led transformation. So you will see a number of use cases today. Some of the demos that will showcase the depth of our transformation capability and hopefully you will walk away with a good understanding of how we get the domain and the transformational capabilities together to improve the value or to improve the efficiencies and operations and add value to our clients. Again clients and so on.

We've discussed in the past the key messages relationship strength. We work with some of the large payers in the U.S. Today we work with seven of the top 10 payers and three or four of these relationships have been around for over 20 years. So if you'd look at the top five clients, the average tenure of those clients is 18 years.

So these are long-term relationships and like I've mentioned in the past not only are they long-term relationships, these relationships continue to grow and we'll show you numbers from the last 12 months on how the top five and the rest of the clients have grown. So very sticky relationships, we've proven to them that we are a great partner. We've added value to them over the years, and they still continue to grow with us and in terms of the scale of operations operate from five geographies, multiple sites within these geographies.

Page | 3

We won't cover a lot about operations today but one thing that underlies everything that we are doing today and that we intend to do tomorrow is the operational strength that we bring. So we've done this work for 25 plus years. We are known as leaders when it comes to delivering on operational metrics and KRAs and that is the foundational trust on which everything else that we do is built. In terms of financials, again we don't intend to get deeper into the financials, but in the Q&A session happy to answer questions, but if you look at our track record we have grown consistently in double-digits over the last five years and our margins have also been consistent in the 24%, 25% EBITDA range.

In terms of recognition, so all of the analysts, the industry analysts who track healthcare operations in the U.S. If you look at all of the reports, you will see that we feature in the top quadrant. So be it Everest, Avasant, HFS, ISG and so on. But within that, I would like to highlight a few of the more recent ones especially as it relates to today's topic which is what does AI mean to our business. If you look at Everest PEAK Matrix, we have been rated in the Leaders Quadrant for intelligent payer operation and the word intelligent is use of transformational technologies like AI and so on.

Similarly ISG has stated us as leader in generative AI services and Business Intelligence Group have taken our use case on Sagility SmarTec Nurse Assist to award us the winner of the Augmented Intelligence Award and you will see a demo of this in action later in this presentation. So this many of you in one-on-one meetings and conferences we have gone through this.

I'm not going to spend a lot of time on this. This is just to say that the statement that I made. One of the pillars that we go to market on is our breadth and depth of our services and this kind of showcases that breadth and depth. If you look at the value chain the first five pillars are all related to payer business, payer and PBM practices. As you can see if you look at the lifecycle of let's say a member, an insured who goes through the payer organisation all the way from member acquisition up until renewal of that member. Everything that the payer needs to do to run their operations we touch almost all of those functions.

We group them logically into these practice areas and then my team that will come after me will dive deeper into some of these practice areas and tell you how the domain depth has helped us build solutions that solve critical business problems for our clients. But member lifecycle, provider lifecycle, claims, clinical payment integrity, these are the practice areas on which all of our solutions are built and underlying all of this is

Page | 4

consulting and technology services. So we do business unit assessments. We do journey mapping.

We do some of the upfront consulting to identify the opportunities with some of our clients and we also have like I said a transformation team where some of the point solutions and some of the Agentic AI deployments we do across all of these practices and like I said you will see some examples of those in the later sessions. So one of the questions that we discussed recently especially at the beginning of this year is all of the changes that are happening in the U.S. healthcare system and how Sagility is navigating those changes and so thought I will take five to 10 minutes to just cover some of the big changes in the U.S. healthcare and how we are well positioned to help our clients through those changes.

Again won't treat all of this and we have covered this in earnings calls in the past, but there have been a slew of changes in the last three to six months. A lot of them just picking up on the One Big Beautiful Bill. There has been funding cuts specifically impacting the Medicaid business. That is likely to play out in the coming months and years.

So we haven't seen the impact of it yet. Most of it will take effect towards the end of the year and the beginning of next calendar year, but it is expected Medicaid heavy plans are expected to lose some members. We will talk about the Medicare not only Medicare but across the board we will talk about the increase in utilisation and the challenges that that poses. So we have a couple of slides which will dive deeper into that and AI.

So today all through today's session we will talk about AI sorry -- Agentic AI and the impact of that on our business and how we plan to capitalise on that. But yes, there are constraints in the U.S. healthcare system when it comes to the use of AI. And we will talk about some of those complexities and the constraints and then explain to you how we plan to leverage the power of AI to increase the efficiency of our operations. ACA subsidy, we have spoken about this in the past and this has played out in the open enrolment season that happened early this year. So there has been a decline in new consumers and there has also been a decline in overall enrolment.

Tariffs not something that impacts us much. Obviously there have been some changes recently, but even before there was not much of an impact to us and we don't expect this to impact us and then the H1B there were questions around that and we have clarified. We have people in single digits with H1B Visa. As all of our employees in the U.S. over 99% of them

Page | 5

are either Green Card holders or citizens and so this is not something that impacts us. So in a nutshell, if you look at most of these changes what it means is two things.

There is membership volatility in many of the plans and two there is a margin compression right and let's double click on the margin compression issue. And this is a representation of the top payers right. Some of them are our clients, some of them are not, but we just wanted to give you what is happening across just not the top payers, but across the industry. The thing that is very common across all of the payers is the increase in what we call the MLR. So if you look at all these graphs and these represent over four years.

The MLRs have gone up from the low 80s to high 80s and in some cases even 90 plus. What that basically tells you is payers are increasingly spending over 90% of their premiums in just paying claims. So essentially left with very little money to run operations and make a profit and so this is the margin compression that you keep seeing in all of the earnings calls of our clients.

All of their CEOs and CFOs if you look at the consistent teams, it talks about higher utilisation how they need to improve their margins. People are talking about this being the trough from which they hope to recover and a lot of them are talking about cost takeout, reducing administrative expenses.

So all of the challenges are around margins and how to improve profitability. Some of them have strategically exited certain markets which are high utilisation markets and so that is caused also membership changes. Some have lost membership consciously, some have gained membership, but ultimately this is all with respect to how to improve their bottom lines. So, the question that many of you asked us is in this context since you guys work with a number of large payers. What does it mean for you?

How do you expect to grow when your clients are going through all of these issues, and we wanted to spend some time and talk about how essentially some of these margin compression issues that our clients face is a tailwind for us and why is that. It's not that everybody can go around saying that it's going to benefit them, but we are uniquely positioned for the following reasons. First, we deal with healthcare operations.

Page | 6

So one of the things that keeps constantly coming up is how are you different from IT services players and we have said that the spend on us is not discretionary spend. Clients have to continue to run their operations and one of the things to improve margin is to reduce the cost of operations and that is where partners like us come into play.

So our domain expertise is one of the main reasons why clients expect us or expect partners like us to be able to reduce cost when it comes to their operations and secondly as I said we worked with some of these clients for 15, 20 years right and so we have proven over a period of time that we are credible that we can scale and we can deliver value right and so when clients want a quick cost reduction, we are the ones who can do it with very low transition cost and timelines because we understand their ecosystem, we understand their processes and we have a very good understanding of their business in some cases even better than some of their own people. And third we have spoken about the transformation capabilities.

A lot of the work that today clients expect us to do is over and beyond just the cost differential between their geography and delivery geography. A lot of it is around how can you further transform a business. How can you take cost out of my operations and all of the capabilities that we have built and some of that that we will demonstrate to you today will show you how we use those capabilities to add value to or take cost out and reduce overall cost of operations for our clients.

And lastly and Roopam will cover this. We are open to multiple business constructs. One of the challenges for clients today is if they need to do transformation they have to invest significantly upfront and some of our business constructs basically, we tell the clients that they don't have to invest significantly at the start. We will make the investments on their behalf and we will straight line the savings for them over a period of time.

So those kinds of constructs we are more open to and we are open to that because we are very confident in our ability to execute on the transformation and show the savings to the clients and so to answer that question if you look at it both from a viewpoint of top five clients as well as the rest of the clients. You can see in the last 12 months which is December 2025 over 2024, the top five grew close to 10%. 9.9% is the growth of our top five and if you look at the rest of the clients it's even more healthier. It's 28% growth.

So across the board both in the top clients and in the rest of the portfolio we have grown in spite of the macro pressures on the industry. So the

Page | 7

cost, the profitability challenges are real and in that environment over the last three, four years we have continued to show sustained growth across our portfolio of clients. So transitioning from there. So, like I said our clients expect and this is not new.

One of the things I wanted to talk to you about is the fact that outsourcing and offshoring gives them a cost advantage is a known thing. So that was the origin of offshoring. If you go back 25 years, the whole premise of offshoring was you deliver it from a lower cost geography and that gives them a big cost reduction. But that's kind of table stakes and it's been table stakes not just now, but it's been table stakes for a number of years now.

So the added value that clients expect from service providers like us is beyond that cost differential what additional value that you can bring to the table and that value being someone who runs operation is how do you reduce my cost of operations. So reducing the cost of operations on an ongoing basis is part of our DNA. So it's not new today. It's been there with us for the last 15 plus years and this is the journey that we've taken in reducing the cost of operations.

If you look at the early days, it was all about process improvement right. So all of you would have heard of the terms Six Sigma and SOP redesign business process reengineering. All of those were things that we did in the early days to essentially look at the processes upstream, downstream, look at some of the handoffs. How do we eliminate unnecessary task? How do we improve efficiency? Ultimately resulting in lower cost of operations. In the last 10 years, 2015 onwards when some of the digital tools became available.

Things like RPA, things like analytics, process mining. These are capabilities that we used in the last 10 years essentially to identify what are the bottlenecks in the process and in some of the simpler rule-based processes can be used RPA bots to improve the efficiency of operations and they have generated a reasonable amount of efficiency.

If you look at typical RPA bots that we've deployed in the last 10 years at a process level they could generate anywhere between 15%, 20% efficiency. So that's something that we've generated in the processes that are amenable to such interventions and those are efficiencies that in some form of fashion we've passed back to our clients. And now in the last 12, 18 months since the road show, we have been talking about this. We have invested heavily in Gen AI and a number of use cases we have

Page | 8

demonstrated to you which are more in nature of if I were to call them more passive uses of Gen AI.

We've embedded them at a process level. They either do, they automate some portion of the task. They help retrieve knowledge based on which our humans have to make decisions and in some cases they've also automated documentation. And all of that we've done making sure that there are enough guardrails in place and so on. So this has been the journey so far, and now when we look at what we've been doing in the last six months or so and into the future. It's to take this to the next level.

But before we do that, I wanted to talk about some of the constraints that the U.S. healthcare industry has, and how as company given our transformational capabilities and our domain expertise. How do we plan to navigate through some of those constraints and still show tremendous value to our clients. If you look at the U.S. healthcare landscape, I want to talk about some of the constraints across these four constructs. The regulatory issues, the clinical judgment required in a lot of the work that we do and also the whole healthcare ecosystem. The number of stakeholders, the relationship between the stakeholders and the complexities of some of those relationships. And then lastly we will talk about fragmented data and legacy systems which is something that I'm assuming a lot of you would have heard from others in other industries as well, but it's amplified in healthcare.

Talking about the regulatory complexity in healthcare. So I won't go through all of these laws and what they mean, but essentially you can understand that healthcare is a highly regulated industry in the U.S. There are number of laws that we need to abide by and some of these laws do post constraints when it comes to implementation of AI and Agentic AI. A simple example the healthcare HIPAA. HIPAA has very strict guardrails around privacy and security. So there is a minimum use standard. So you can only use the data that's required to complete the task and not overuse. So obviously if you're using AI and Agentic AI, in some cases you can't give access to all of the data, because the law prohibits that.

And when we talk about clinical, we'll also talk about places where there are guidelines and laws today where some of the medical decisions cannot be done by an autonomous agent. What this means overall is when we try and implement AI and Agentic AI in healthcare. You got to be careful that what you do is auditable, what you do is explainable and what you do is compliant. So what that means is it limits how much you can do in an autonomous way and we'll show you some examples where we've tried to blend what power we can get from an Agentic AI implementation

Page | 9

along with human in the loop to be able to accomplish greater efficiency for our clients.

Next, when you look at clinical judgment. There are number of factors and I'll touch upon a couple of them. One, we spoke about the regulatory aspect of it in which a lot of the clinical decisions things that impact the care cannot be taken autonomously. A human always needs to be involved in making those decisions. So that's one constraint, but even otherwise healthcare is not something that in some cases can be done autonomously, because for example and Krithika will talk about this. In a clinical care gap it is seen that the member or the patient doesn't understand that there is a gap, but it's some societal barriers, it's some behavioural reasons because of which the member or the insured isn't taking the right action. And in those cases you need a human to be able to understand and do the right nudging so that those gaps can be closed.

And more importantly one of the bigger concerns when it comes to AI is, is there going to be a bias in terms of parity to some of the behavioural health and other benefits that members are supposed to have. And so if you look at regulations, we need to make sure that there is no bias that violates any parity rules.

Talking about complex stakeholder relationships. We've discussed in the past about the U.S. healthcare ecosystem and how there is an interplay between payers, providers, members, employers. So all of them kind of come together and the incentives of all these different stakeholders are diverse. And more importantly the complexity is also in the number of contractual and payment related relationships.

When we talk of an insurance company, it's not that they sell one single plan. So every state, every county, the variations of the plans could be different. The benefits could be different. Similarly, every provider relationship, every provider contract is different. There are nuances in those contracts. So much so that one of the acquisitions we did DC on payment integrity. The service there is just to reprice the contracts. I mean these are contracts that have been loaded onto the system. Auto adjudication rates are in the high 80s and 90s.

In spite of that we have an opportunity to go back and review the payments that have been made against the contract, and we find more than enough opportunity where the payer has actually overpaid, because the intent of the contract wasn't really established when they set it up in the system to make those auto adjudications. So those kinds of

Page | 10

complexities in terms of the number of permutation combination and what needs to get paid is immense.

And also when it comes to the member or the patient there is very limited. I mean this is healthcare is one of the products where the people who actually buy those products have very little understanding of how it works. Even as a consumer, members have fairly limited understanding of how this works and that creates the additional level of complexity.

What this means is as practitioners as people who run healthcare operation we need to have a very strong understanding of the ecosystem to be able to precisely deploy AI. And you'll see in some of the examples that we showed today how that understanding of the total healthcare ecosystem is very important to be able to do this.

And lastly, like I said you would have heard this from others as well. The fragmented data and legacy technology infrastructure, but healthcare and we've discussed this in the past and most of you know healthcare is from -- as an industry is quite behind in terms of modernising some of these platforms. Today if see a lot of these platforms are outdated. A lot of the information across the ecosystem be it membership, billing, claims. They all act in; they are all in siloed systems. There is very little exchange of this data and what it also like I said in because of regulatory constraints sometimes there is also restricted access across these silos, because you need to get consents before you can go across the silos and get the data.

And so what that poses, the challenge that it poses is that an Agentic AI doesn't have access to all of the data in the enterprise. And so you will see some of the examples that we demonstrate today. With those constraints we can still implement an AI based solution, which works with these constraints and still delivers value to the clients. What it means as a practitioner for us is the reason, the knowledge required to understand the client's ecosystem so that we know precisely how to deploy this AI.

Like I said, the question that all of you ask is with all of this advances in AI, why are your clients not adopting it immediately right. And hopefully some of the complexities that we discussed are basically the constraints that our clients and we need to work through to be able to get the power of AI. But we believe that given our history and given our domain expertise we are uniquely positioned to help our clients because of the following reasons.

Page | 11

Like I said, and you will see today we've designed processes in which human is always in the loop and we can use the power of both AI and the human in the loop to deliver much better outcomes for our clients. You will also see the deep clinical and domain knowledge that's required to deliver on some of these promises. And like I said we are very flexible, I mean, even though today a lot of our contracts are on FTE or transaction basis. We are very open at this point in time with clients on multiple pricing constructs and we are willing to take risks to be able to deliver those efficiencies to our clients.

And finally our proprietary tech and our ability to integrate AI into the process workflow is something that will separate us from others. So it's a combination of all of these capabilities that gives us the confidence that we are better positioned to take advantage of the technology.

And so the way we see the future evolving is rather than using AI passively in a specific process we are at the stage where we are embedding AI in our end-to-end workflow. So you will see demonstration of that today. We've created a whole set of what we call SmarTec agents. One, I'll talk about the internal one, but essentially in the work that we do for our clients both on the administrative side on the clinical side these SmarTec agents are utilised in complex workflow. So we can automate a complex workflow from end-to-end and use agents and humans appropriately in the workflow cycle to deliver the value that the clients are expecting us to deliver. And you will see specific examples of that today.

And then the final holy grail where how do I -- if I'm able to overcome some of the data constraints and the legacy system constraints. How can I even increase the power of AI by having AI do the reasoning and the decisionmaking and that's where our Synchrony series of solutions come into play. And we'll give you some examples of that today. So essentially think of it as an offering in which we take a life cycle end-to-end. For example the Medicare Synchrony product takes the membership enrolment life cycle all the way from plan design to enrolment and billing and within this life cycle how do we bring partner platforms and create an orchestrated agentic workflow on top of that. So that within that lifecycle the agents can make autonomous decisions with appropriate human intervention but be at a much higher level than what it could do in a legacy environment.

So that's where we think the future is and you will see today that some of the partners that we are working with, some of the partner products that we are working with. So in that new ecosystem both with the partners and with our expertise will create that agentic workflow like I said with a much

Page | 12

higher level of autonomous decision-making. Then what we can do in the previous era. So those are the two things where we think the future is headed and we wanted to give you a taste of that the coming presentations.

The subsequent presentations is like I said going to focus more on how do we execute on the strategy. And like I said, deep domain expertise is a key prerequisite for execution of the strategy, and so Roopam, Srikanth and Krithika will touch upon how we bring that expertise to the table and how that expertise helps us deliver on the promise of AI. So you will see some of the new business constructs. You will see how our understanding of payer and provider ecosystem gives us the depth. And like I said, you will also see some examples of what we mean by Synchrony and what value it brings to the client.

Second is, we'll also have presentations on our tech-led transformation so that's going to be a lot of demos on how we use Agentic AI and what is the value that it delivers to our clients. And you'll also get to see how we're using the partnership ecosystem to deliver on that value. And towards the end Chris will talk about how we are taking all of these capabilities to market.

Like I said our large clients as well as the rest of the clients continue to grow at a healthy pace, and so where is that growth going to come from? What are the solutions that we plan to take to that market? And then the strategy of penetrating the mid and small market is something that we are very bullish on and what kinds of solutions appeal to that clientele and how we plan to do that and also some of the new services that we are working on. So all of this we will cover in our last section.

So at the end of the day you will see how the domain expertise, how the transformational capabilities and taking all of this in a very focused way to the market. How that's going to power our growth in the coming years. So with that there is 10 minutes. I can possibly take a couple of questions and then pass it on to the next speaker.

Atul Mehra:

Ramesh Gopalan:

Yeah, hi. Thanks for hosting this. Just one question, so the biggest apprehension today like we all know is around AI and so on and so forth. So can you give an example. So you spoke about how difficult complex things are. But could you explain it with an example how AI could…?

That's in the next session. You will see some live examples of areas where we've deployed and you will see a…

Page | 13

Atul Mehra:

Ramesh Gopalan:

Deployed, but also in the sense that, like what is it in our business that a way is AI proof in some sense or how things not be like completely like our business model may be not impacted to the fullest extent or to any major extent because of AI. If you can like walk through some real life example. Thank you.

I think Roopam who comes after me is going to give you a real-life example. So I don't want to steal his thunder. But the broad response to your question is, we are looking at AI as a force multiplier. We are not looking at AI as a disruptor. Of course there will be disruption like I said. Generating efficiencies and operations and passing it back to the clients is part of our DNA. So that's something that goes, that's an ongoing part of our business, and so it's happened in the past. It will happen in the future. But today with AI and Agentic AI, we think it's a capability that will really open up additional opportunities for us and given in the context of the market, which is the cost pressures and the margin pressures that a lot of our clients’ face.

We believe that this is essentially going to help us even grow faster than we have in the past. It's a question of how we embrace it, how we deploy it and how thoughtfully we deploy it keeping into account all of the constraints that we spoke about. And I think given the position that we are in from a domain depth, we've done this for 25 years. We understand the ecosystem, and so we believe we are very uniquely positioned to help our clients in this journey. So we are looking at AI and Gen AI as a force multiplier and not something to be concerned about and that's the message that I wanted to leave with all of you.

As you see through the rest of the presentation, we are essentially embracing this as a tool that can reduce cost of operations for our clients and given this in the current environment where most of them are struggling for profitability. How do we use this to add value to our clients, make the relationship stickier and continue to grow with them and places where we will grow and how we plan to grow is what Roopam and the rest of the team will talk to you about.

Pallavi Deshpande:

Ramesh Gopalan:

What we're saying there is no deflation of impact that what we see in the IT services space that is not something we see in this space.

Again, right at any efficiency could be and we'll talk about it maybe in more detail after all the presentations right. Adding value to our clients has two aspects to it. This one way to kind of impact the invoice that I build to my clients, which is generating the efficiency in the operations that I do

Page | 14

for clients, but the bigger impact is what we say cost avoidance for the client.

So if you see a number of examples that we'll show it not only means that the only way to add value to my clients is not just by reducing my invoice. Yes, that's a component and AI will impact that, but the bigger value that AI can add to my client is how can I help my clients avoid cost that they would have otherwise incurred? So that's the bigger play. And so to your question at the micro level, at an SOW level if I generate efficiencies will it impact the revenue I generate? Yes, but as you can see from the rest of the presentation it also opens up opportunities for us to enlarge the scope of the work that we do.

Look at a process end-to-end whereas we were operating in silos before and how that will increase the pie. So at a micro level, yes, there will be some reduction, but it also opens up an additional stream of opportunities for us.

Participant:

Ramesh Gopalan:

Just wanted to understand, so what AI has done is the provider have always been disaggregated and may not have access to the best technologies as the peers have. But in AI, we are seeing multiple companies trying to provide these unique solutions, autonomous coding or better transcription, which leads to better claims in the first place. So what AI is likely to do is improve the provider's journey. And in that context, how do we add that better value to keep the payer more competitive? Because it almost looks like much of the AI is coming after the payer. In that context, how would you add significant value going forward either using AI or maybe whichever way possible.

It doesn't change the answer right. Everything that we spoke about so far is to make the payer more competitive, is to reduce the cost of operations. And we'll also show you an example how we bring the knowledge of having worked with providers and payers and how we bring value to that interaction between a payer and a provider as well.

So there's an example where we talk of a HealthBridge Connect, which is basically a connect between payers and providers to reduce the cost of both payers and providers. Because every time for example a simple example a provider builds a claim and doesn't get paid on time or doesn't get paid the amount that they think they should get paid. It causes friction. It causes an interaction. It adds to the cost. How do we work since we know what are the reasons why that friction exists. How do we work with both parties to reduce cost for both parties. These are friction points

Page | 15

which if we use the right intervention and technology, we can reduce cost for both parties.

And so those are some of the interventions that we've done with some of our large clients which have been very useful. But broadly everything that you will see and hear today is essentially how to reduce cost of operations for our payers right and large part of that is the relationship the payer has with the provider. Some of the friction points in those interactions and how to reduce that friction.

Participant:

Ramesh Gopalan:

Just a quick question. I understand that the cost of operations reduce to some extent. So in that way, do you believe the entire services by both outsource or in-house done by the payers, and that size will reduce, but you will end up getting a higher part of that as outsourced thing, what does the reduced size of the services is? Because some of the AI would also help them automate their own internal processes right now. Some of them will help you automate the process you're providing them. Is there a leakage of your processes going back to them in-house and maybe more opportunity of their in-house process coming back to because of the AI you generate. If you can give them that numerical answer that would just be.

Yeah, so before we get into the details, one of the things that I've shared with lot of you is the macro status on outsourcing penetration. So if you remember doing the road show and subsequently the slides we showed. Payer as a whole the outsourcing penetration is in the low 20s. So when all of us start saying oh if this is automated what's going to happen is payer going to take the work in-house. It's the other way around. There is still 70% plus work that is done internally by the payers today right.

So the efficiency is going to be generated in both parts of the pie. Both in the work that we do currently for the payers, but more importantly the large portion of the work that they still currently do in-house. And some of the business constructs and some of the examples that you'll see today is exactly to address that right. How do we use the power of AI to increase like I said, to increase the opportunity size. We might be doing just one function for the client, but because today if I can combine the upstream and downstream and generate a higher value, the client is more than happy to let us do that.

So while yes the work that I do might generate efficiencies and disrupt my revenue there, but it also increases the size of the pie. So you'll see some examples of that.

Page | 16

Namit Arora:

Ramesh Gopalan:

Hi, Namit here from INDGrowth. Yeah, my question was around within your own organisation, how are you leveraging AI within Sagility as an organisation? And secondly, how are you using AI to develop some AI native offerings for the future rather than sort of looking at the current? Thank you.

I mean we didn't talk about that. He is chomping at the bits to talk about what we are doing internally in the organisation. So, yeah I mean one part of AI and the part that we are going to discuss today is the offerings that we take to clients. But AI is something that we are also embracing internally as an organisation. Everything that we do today right starting from hiring to training and everything is going to be completely reimagined using AI and all of that is in the works.

So, we believe in the next 12 to 18 months that's going to have a big impact on the cost of doing business for us and we'll talk about that more in subsequent earnings calls and so on. But there are a number of initiatives underway to make our operations much more efficient because today as you know a large part of what we do is hire people and because we hire them in multiple geographies where there is very little understanding of U.S. healthcare. A lot of time and investment goes in hiring and training those people and how can AI impact that significantly. So those initiatives are also under play and if you have time towards the end, I'll ask Anand to give you some more examples.

Namit Arora:

Ramesh Gopalan:

Roopam Narayan:

Thank you. Thank you again for hosting this and having the entire senior team here. Thank you.

Thank you. We will move on to the next session. Like I said, if we have time at the end of each session, we will take a few questions, but we will have about 45 minutes at the end to take other questions. So thank you.

Thank you, Ramesh. Hi, good afternoon, good evening. My name is Roopam Narayan, I take care of Practice & Solutions. I'm based out of Denver, Colorado. Some of the questions that you asked and probably will get answered in this session itself. So one of the things about looking at to AI or any of these tools is that customers can implement it. Service providers like us can implement it. Who can implement it faster? Who can give results and ROI faster and better?

And who can commit to an ROI. I mean these are the questions that will make the decision of whether we grab more business out of the client or whether we just lose part of our own revenue. So that's where the entire session that I'm going to talk about will come into this. So we have our

Page | 17

traditional methods of growing. We are growing in our existing as well as new clients.

But we have specifically focused on outcome-based offerings leveraging AI primarily to grow much faster and grab market share which was not available to us. It was not even part of the addressable market for most of the outsourcing vendors in the past. And that comes to us because we are actually an AI first organisation. So I'll cover three and I have examples for this one and my colleagues Sri and Krithika, they'll come and give some additional details and these are with numbers so you can probably see the power of AI and how these impact our clients business and how it impacts us. So it's a win-win for both the organisations.

So Managed Services Structure deals, you probably would have seen a lot on the IT services side. It hasn't happened much on the business process side. So let's start from there. And in this part, like any other IT services, managed services deal, client outcomes are very clear that they are looking at cost savings committed, number one. Number two, they are looking at speed to value.

So you don't have to really wait for two years, three years of your automation projects to kick in and therefore you get the refund. They are also looking at zero or no CapEx, no upfront spending. So that's really needs to be part of the deal. So that's the benefit. And then the differentiation comes in operations specifically which touches their members, providers, external stakeholders directly and that's why it's not discretionary is that they do not want a black box outsourcing model.

So a lot of part that builds is really the trust and that comes through incumbency. That comes through a very strong delivery track record that we have these clients. So typically most of our managed services proposals that have gone in and where we have signed up or we are going to sign now in next maybe months to quarters are really with the existing customers. And that's the benefit that you will see from the client outcomes.

On our side, like we said, is a scope 70 plus percent scope which was previously not available to us. It was retained by the client. That becomes open to us because managed services always end-to-end. So we will take the entire enterprise volume and do an end-to-end managed service. That comes in and most of the scope issues not coming to us was onshore.

Obviously at a higher per FTE cost rate and therefore the automation has much higher benefits there. ROI is much higher there. So that's part you

Page | 18

will see from that part of it. We get the control on automation projects. So far we have been working with the client in partnerships and we go and give some suggestions and we provide the automation. Their IT team will come in best efforts basis. They will go deploy the automation in the production environment and then it will take another three to six months before value is realised.

But here in this case, as part of the contractual agreements, we have secured control over the automation projects including that with the Gen AI. So that's the second part that is good for us. And this breaks the linearity. There is no discussion now on transactions or FTEs. This is actually getting into base fee and ARCs and Rooks are more aligned with their revenue which is membership. So if you acquire more membership, we get more revenue. If you lose membership of course, still a floor. We are willing to let go of some part of the revenue because that's how this entire model works.

So this is a very powerful one and I have a live example of one of the scope which I will cover in the next slide. Coming further, now we are talking of getting into the mid-market. We have some clients. These are the clients and I think there was a question on payers are usually the businesses aggregated at top with the payers and then provider is not completely fragmented and disaggregated business like I think you brought that up.

Even in the payer side, if you leave out the top 10 payers, the subscale payers have tough time competing in the market. And that's because not the cost of operation, which is still fairly linear, it's cost of underlying technology. That's where they are looking at partners who can come in, who can bundle in technology and give a Synchrony, what we call it as an AI led solution, but it also brings in platforms so that data is not in five different platforms and there is lot of M&A activity which happens in those smaller plants. So it's that they have three plans which have come together and now three different claims platforms. So how do you address that? So that's where our Synchrony services come in.

Where we are coming not just for the operations but including the tech platforms underneath it. Some models may not require us to own the licenses. Clients can own the license, but we still have the IT production support, IT support everything is part of our offering. What's the other value from the client point of the area? Speed to Value. When client like that which is stretched for resources, they will take years to implement or integrated multiple platforms.

Page | 19

When we do it, we are going to convert those years to less than a year, nine months, 10 months and platform is alive. ROI is much faster, much better, realise that's something which comes in. And then what we are doing is the biggest cost is really technology and technology is really upkeep and making it compliant. Your CMS changes, Interop came in, No Surprises Act came in, pricing transparency came in.

I mean how many IT technology are you going to keep on adding in every module, every platform. It's extremely difficult. But we do it on our platforms. It serves three clients, four clients, five clients, shared service. So that's something which we can deliver for these people and it becomes the platform becomes future proofing for these clients.

And if you see that those are the client benefits, very similar benefits for us. We get higher penetration in the market with a higher wallet share. Because now we are not just talking of offshore claims processing, we are talking of claims, conflict, claims payments, provider data management, provider calls everything put together is one bundle. EDI, file exchanges, IT production support everything is part of that one Synchrony Claims Lifecycle the way we call it. It is all covered in one process.

We take end-to-end, we deliver it. So that's where you get the first two bullets of our penetration in mid-market and high wallet share with the same customers and it's a very sticky model. Like I said is that if you see the M&A activities which have happened so many plans came with Medicare and the moment the reimbursement rate dropped, increase dropped, they have either shutdown, they are not in the Medicare market anymore or they got bought by a larger plan. And we have ourselves lost business. I mean we lost a care management business in a similar plan because they got acquired by a larger plan.

But when you have a platform-based business, the acquirer cannot change the platform overnight. So it basically gives you a sticky five, seven-year kind of contract. So that's the other part of benefit that we get out of the Synchrony Lifecycle Operations.

And lastly with the MLR becoming such a big problem. A lot of growth has happened in the MLR. You probably would have heard not from us but from even others is that clinical business is actually growing faster than the rest. And the reason is simple that if your medical cost is going up something has to be done. There are other actions of renegotiating provider contracts etc or changing the benefits, making more cost share go to the member that insurance company doesn't have the 100% liability of those claims. So those are adjustments on your benefit products

Page | 20

provider contracts, but a lot is about how can you change your health of the population.

So they are not consuming high cost expensive items in your healthcare and therefore get the benefits. MLR reduction services, we differentiate from managed services. Look if somebody gave me a UM process saying that it's a 6 million process, can you run it in 5 million. It's a managed services. But when we are talking of a 6 million spend and the plan is saying is that I am getting a medical cost impact of 18 million, can this become 30 million.

So then the discussion is no longer of 6 million versus 5 million. It's a question what is a medical cost, what is a medical impact. So that is what we are focusing on that and Krithika will cover actually with the journey of a member patient you can call it and how it works in MLR reduction part of this one. So that's the three areas that you will see us focusing on and we have some early wins in these areas. We have some good pipeline.

We have tested these offerings. All of them the lynchpin, the bedrock for this is very strong delivery which is still our strength and a very strong domain led technology that is again our strength which are the part which goes inside these offerings. I have one, one slide each one for these before we get into individual cases and we will drill down deeper. So next slide is really one of the constructs. It is close to the numbers that we have negotiated.

Now this is a process which is costing 25 million to the client. It is a single LOB or one of the LOBs doesn't have the member lifecycle. It does not get into enrolment and that side of this house. But this is only claims with some provider lifecycle. So provider data management the claims calls that come in because of the provider claims etc is this one.

Like I said we normally talk of 6 million is our revenue for last eight-10 years. Give or take with membership increase which happens because of population and automation benefits that go like Ramesh said that it goes back to the customer. It has been stable around that. Maybe it will reduce for Gen AI because the productivity will be higher than the membership growth that we will get.

But it's 6 million there but you see the pie we don't have. 13.5 is retained with just about 15%, 20% members who are so called the restricted groups which cannot be offshored. So the processing happens onshore client retains it. There are a couple of other vendors in the 5.5 million. But

Page | 21

when you're doing a full end-to-end because any Gen AI deployment, any such deployment is an enterprise level.

When the client did their ROI mapping and everything they said, we can reduce this and give you 5 million starting day one. Okay. Because we are using the same technology that the client allowed us to do. So our technology team is at parity with theirs. They have other priorities. They have other projects. By the time they come and do this, it's going to take them time. They will take them longer. We can shorten that cycle and give them the benefit from day one and do a five year deal.

So think from our perspective 6 million possibly declining now becomes instead of 30 million over five year or less. Now is a 100 million for us. So that's a 3.5x growth for us in terms of revenue. And from the client point of view it's no brainer. Okay.

It's not your risk not your investment. The party that you have done is already doing most of your volume which is us. The delivery is working better than anybody else who can deliver. And that's how we are getting this kind of a business. It has a direct impact on our revenue margins and from the client point of view upfront savings. So that's really the kind of benefit that a managed services model that. This slide is a little bit busy. I'll not get into lot of this one. Just to explain that the multiple causes of medical cost increase. That's inevitable.

Okay, you have GLP drugs, you have high consumption on behavioural health, you have an elderly population, you have a backlash that we have seen in the market against claims denial or authorisation denial for services. So, all that points to things happening where medical cost will increase. The way to contain that cost is to have your members take the preventative care in advance and you cannot go chase every member. So how do you identify the right membership chase them and between the two sides, one is control the utilisation so that you don't have the waste. That is one part of the offering which is very important.

The other side is make your population healthy. So that and it kind of feeds together into each other and that's what we have developed using our AI framework and analytics framework and that's what's giving us and when Krithika comes in, she'll take a patient journey and show you how this works and how we had not done the intervention, what was the impact? So that's really from and to situation that we have mapped out and shown this one.

Page | 22

All the discussion then no longer happens on I'm spending $10 million on my Care Management and I have to bring it down to $9 million. Fine you can bring that down. But is the $10 million giving you $15 million impact or is that giving you $30 million impact. There the difference is $20 million, not $1 million worth of saving. So, when we are having discussions in MLR Reduction it's about the impact rather than just getting 10% saving on a given process like a managed services model.

So, this is a very powerful model that will get into. In many cases, what we have started finding now is that clients are actually adding the cost to come with new programs in Care Management, because it's likely to have 3x, 4x, 5x ROI, and we are the ones who will be running those programs. So, it's a new market for us. So that's where we are saying is that market is actually expanding for us including AI platforms and not reducing. So, that's a very different statement that you will hear from us compared to many other vendors.

If we go to the next slide and get into the Synchrony services. You will find that we have put together a series of services that are there in terms of a Payer life cycle that we had put in as the first slide. Now in this Synchrony services, enrolment and billing is one part of the membership life cycle. Complex enrolments are in Medicare and ACA plan. So, we have fully flesh out Medicare ACAs under this one. When Sri comes in, he will talk about a Medicare life cycle under the enrolment building.

We have Claims Life Cycle. The previous slide where I showed the managed services that was a managed services for claims, but we are in discussion on a Synchrony where we are bringing in our platform with a partner for the Claims Lifecycle. Utilisation Management again we have couple of very strong partners and right now we are I think Krithika will probably have the right numbers 21 identified plans where our Synchrony Utilisation Management with the partners is being targeted for. And these are mid-market. Mid-market I will define as less than a million members. These are not very small, but they are still smaller than the plans that we typically handle today.

SmartStep Aging in Place is our award-winning platform-based service. Long-term careers are already using this. So far, it's been used by longterm insurance companies primarily because their objective is not to have a higher percentage of their members go and use long-term care facility because it costs them money. So, they have been using it. But now it's come as part of offering of that you can do post-acute care and keep the member at home and well at home, taking care at home so that the member does not go back to some skilled nursing home facility where

Page | 23

again there is big spending for the health plan. So, you will find those kinds of solutions which have direct impact to outcomes of our customers. And therefore, they are willing to pay us our right share of revenue and margins. And none of these are linked to actual transactions or FTEs or something like this, because these are all outcomes.

And these are customers who are tested with, or we already have a strong revenue coming from them already. PI services, again, it's a platformbased service. Ramesh pointed out that we can look into the old contracts and come back with that where the client or payer client overpaid and therefore there is recovery. We recover that money and we get contingency based payments for it. So again, it's contingency based model. There's no FTE based linearity. AI driven, all the contract loading etc. I think somebody asked about internal use of AI. These platforms are not being sold to the customer. So, all efficiency that we have to load hundreds and thousands of contracts and thousands of fee schedules is all Gen AI driven.

So, it all brings everything together and we get much faster results on our PI than most of our competitors, and therefore it improves our margin and revenue, because the faster you recover that's your revenue. So that's something that you will find.

A&G is something that you will probably see in our technology demo on appeals on how it's completely been. Wherever we can automate components of appeal of course there is human in the loop. That's again part of the solution that you will see there.

RCM is in the provider side that's the offering that's again platform based. So, all these are AI driven with an underlying platform that we have, or we bring in with our partners. The client doesn't have to pay for platform. Client doesn't need to pay for our FTEs. Clients are actually paying us for outcomes. So that's really the kind of positioning that you will have with each of the Synchrony services, which differentiates us, which also breaks the linearity with the number of FTEs or transactions or any of these models. And Gen AI here is our friend. So, think like that. So, it's basically enabling us to compete in the market, grab market share and really improve our margins. That's really the three objectives the way I'll put out of the use of AI when it comes to our business.

So this is where I just wanted to leave that these are the three kinds of offerings which are all outcome based, all Gen AI based and if you now go back to the questions that were there, you will understand that Gen AI is

Page | 24

actually what we are leveraging to grow our business, grow our margins by giving even higher impact and higher ROI to our customers.

We completely understand if you don't do anything what the expectation is that your revenue is going to drop. Some business will get pulled back, because the client has already automated much faster. All those things will happen. But it will not happen if you have these outcome-based services which are actually giving ROI and speed to value which is better than what most of the plans can do themselves internally. So, I will leave it that.

The next two sessions will get into the core admin side which is the payerprovider synergy that will be with Sri. After that the clinical portion which Krithika will explain clinical gets a little bit too technical. So, we are trying to take it through a member patient journey in terms of without generative AI based solutions that we had, what was happening, how we improve the UM experience, which is to avoid the waste and get the right service authorised for the patient.

And then the next phase was how post service, how we can take care of the same patient with our same Gen AI-based platform services and ensure that you do not have readmissions, you do not have problems of somebody cannot get taken care of after a surgery at home and therefore has to go back to some nursing facility. So, all that you will hear through a patient journey and these are I mean of course in all HIPAA and masking our customer data we kind of taken the information from there and created a journey map for that.

So, hopefully, in this session we could convey what is happening in the market and how our offerings and the growth numbers that you have seen this year which will continue in the later on years. How that is getting delivered using our domain expertise and technology.

Okay, I also have about six minutes, maybe one or two questions I can take, and then we will move to Sri.

Shreesti Rastogi:

Hi, this is Shreesti here from Canara Robeco AMC. Just wanted to understand on the managed services piece and we spoke about how from $6 million we became a $20 million deal. Did not understand exactly what happened because we said that 13.5 million is unrestricted. So, from unrestricted to actually getting awarded to us, it was difficult to comprehend. That was one.

Secondly, when we speak about AI life cycles and managed services deals that we are now incrementally working on. So first of all, what

Page | 25

percentage is these in terms of revenue and how does it help us in terms of protecting our existing business? If you could quantify that that would be more helpful for us to understand.

Roopam Narayan:

Okay. So, let's go with the first one. First of all, the premise is Gen AI is across the board. It can do restricted as well as non-restricted business, right. So, it doesn't matter to Gen AI agents. Autonomous agents can handle restricted business and non-restricted. So, when you are deploying Gen AI solutions, let's say I am just giving an example of claims match with provider data, which is causing all the paints and you have to resolve this. It's a Gen AI solution from us which is going to do that. That works in the restricted business, it works in the business that we have, it works in the business that other vendors also have the volume.

But for us to bring our technology and give it their two methods either we go sell the technology, charge per transaction to the customer that's another method or there is something like that which is a traditional model. The other is which we ask the client that give us your entire onshore, your other vendors business and our business, anyway we have it, okay. You can stay out of it. Now you don't have to pay for any of that, because we are deploying our technology. We are deploying Gen AI agents. We are automating. And therefore, the entire revenue. So out of the 20 million we have onshore resources. So, the fallouts after Gen AI is still being managed by us. But it will be our agents not the clients’ agents.

So, let's say because Gen AI agents will become our resources not the client. So, client does not have any resources after this. It's all Sagility resources who are working on the system, okay. So, I hope you got that that how that business has come to us. Onshore offshore, okay. So, when it gets restricted for onshore, so traditional methods now you also if you go back to the history the way Ramesh was explaining. There is no value for onshore business to be outsourced because there is no labour arbitrage there.

So, if you see most of the large plans and outsourcing the onshore business has not been outsourced. This method allows us to take the onshore business put the entire end-to-end cost. Put Gen AI on top of it and deliver value for it. Good question. I hope this was a good clarification on how that $6 million became $20 million for us, okay.

Your second question. I'm sorry, can you repeat that one?

Shreesti Rastogi:

Yeah, I mean, this is my understanding, but what we are talking right now is what we are doing right now to protect our revenue stream, right?

Page | 26

Managed Services, Synchrony, AI this is not what we're doing historically. So just the question is that what percentage of revenue or deals are like because of this that is one? And secondly how is this helping us in protecting our existing revenue if you can quantify that?

Roopam Narayan:

So, let me once again and I think Ramesh covered this in his talk also is we'll not call it protecting the revenue. There is no protection of revenue against Gen AI. It will erode. Automation will always erode the revenue. The solution to that is what can we do using Gen AI which can give our clients higher value than what they can do internally with Gen AI and then comes these kind of constructs.

And your point is extremely important, because most of the plants are themselves under pressure right now. They don't have CapEx dollars. They don't have lot of investment. So, they are looking at vendors who will bring investment and deliver outcomes for next few years. So, we are in the right place there. We are the incumbent. We have shown this one. We have strong deliveries from technology. We are willing to take that risk, and we are willing to follow their technology guardrails, because everything that we are doing is following the client technology guardrails and we are giving the solution based on that.

So, from all aspects at UCB we have covered, the risk mitigation plans. We've covered their technology risk mitigation plans. We have covered their technology risk come with a financial model that works for them in the year when they need the savings. They need it now, okay. And we are no different from what they would have done anyway with their own technology teams in terms of bringing Gen AI and other things. So that's the part of it.

So, I'll say is that the thought process is not about protecting the revenue, it's about growing the revenue using Gen AI. I mean that's the way I'll put it. Protect the revenue would have been that if we didn't do anything how do I protect and how do I make sure that my volume does not get impacted by Gen AI, we are not in that business. I mean we all for creating more efficiency for our clients and faster. And it's all about giving or showing the impact and delivering the impact which will be better than any other vendor or even their own internal teams and therefore demanding that we get the higher share of business.

Participant:

So, historically, when the IT companies used to take the in-house business from the customer, they used to take the resources also from the customer, that their manpower will go to the IT company and then efficiency again will come to the IT companies. When we are offering

Page | 27

these services to the client, are we saying give us your manpower also, we'll take end-to-end solution for you and we will take care of your manpower also. How we use them? Efficiency will be on us. Is that how we are working?

Roopam Narayan:

That's deal specific. It's not we are against it. It's not that we are for it. So let me just put it this way, rebadging of resources is always on the cards. But it's not the driver for this. Driver is Gen AI, driver is productivity. And we don't actually need because we are doing the same work from offshore. So, we don't have a gap. So many times, IT companies do it, because and we may have to do that if we went for a completely unknown business. There's no transition, so how do I go and run this. So then rebadging becomes a necessity.

In our case and specifically in the case I am rebadging is not a necessity. But it's at the same time it's an optional service. Obviously we're open to that but we're not forced to, because we have gaps in delivery or something like that. We don't have such situations here. I think I'll have to move because our time is up, but if there is something you raise the hand. Okay and your question please. Yeah.

Participant: Hi, I just wanted to know how does your client evaluate you versus the other companies who are also doing something similar? How do they evaluate the cost benefit what you are providing versus what any other competitor will provide and how do they ensure that efficacy of the tools are working?

Roopam Narayan:

Both these questions will take it in the Q&A, and they are similar from the ROI and competitive positioning point of view. So, we will take it from there. Thank you. Thanks a lot.

Srikanth L:

Hi, Srikanth, part of the practice organisation. I manage the core admin side of the work. Personal safe harbour statement, we might get a little too technical, especially me and my colleague Krithika when she comes in. So please in case if there are any clarification, happy to do that a little later.

So, what I am going to cover, I think Roopam and Ramesh talked about the different constructs of what we are doing in terms of adding value to your client. My focus would be on the Managed Services portion that Roopam talked about and the Synchrony services and taking examples of how will we do and how do we deliver this. We will talk about some examples, will talk about some of the solutions so that you are able to see that.

Page | 28

Okay, first example of complexity of the problem that we have. Typically, if you look at a payer and a provider journey, it's very, very siloed. So, you have different organisations within the provider teams who are talking to the payer organisations and different departments. It could be right from contracting, there is a separate entity which deals with all the documents, manages the relationship with the payer organisation. If you are looking at somebody on the claims side that's a separate team.

Even in the payer side or in the provider side these two entities have different systems, these two entities have different SOPs, these two entities have different guidelines. Normally, they don't talk to each other. The advantage we have is that with all the work that we have done over the last 25 years we understand each and every bit of these lines of businesses for both on the payer side, as well as looking at from a provider perspective also as to, hey, why are the friction points happening.

Industry research shows that the payer and the provider friction points are costing close to around 40 billion worth of rework, repeat calls, repeat transaction, repeat claim submission which is impacting the overall administrative cost of a payer or a provider organisation. For example, if a claim needs to be paid accurately you want to make sure that your contracting details that need to be provided are very accurate. If you don't do that upstream if you have any mistake on the contracting side, it impacts all the downstream problems that you face in either claims or appeals or Care Management and so on. That is the complexity.

You were asking about, I think it was your question. What are the examples of where we think it's lot more difficult. It is difficult but that's also the opportunity. What's the opportunity historically what we did with some of our clients is we looked at impacting all the process improvement opportunities or point solutions which are there in these two boxes that I'm showing here. All those were hey we will do point solutions for contracting. We will do point solutions for claims. We will give efficiency of say 10% to 15% on the claims volume of work that we are doing. We will save another 5% because we are doing contracting work that is how traditionally we went about doing all the value addition and the value delivery that we did historically.

Fast forward now, what this has meant is for some of our mid-market clients etc where they have not been able to bring this all together because of our domain expertise, because of we having the view of all the different levers that needs to be brought in to make sure that we are able to give the end-to-end value, we have been able to save a lot more higher value dollars for our clients and we also have proof points.

Page | 29

One more example I will take is for the same client that Roopam talked about previously where about the 6 million to 20 billion. Around 18 months back I think there was one more question how do the client see our value delivery compared to the others. For that particular client the client said that hey Sagility, we have a problem of looking at end-to-end processes. We don't have a view. Why don't you become the convener and create the innovation forum of bringing all the different other vendor partners of ours and make sure that you drive that innovation.

And that innovation idea that included many of our competitors, we were the convener, we were the champion of bringing all those solutions together. We did the entire mapping similar exercise looked at what are the value drivers, what are the important aspects that needs to be done, and we created an overall enterprise pipeline of projects worth close to around 45 million. And in that we drove some of the changes in terms of hey, what are the automation opportunities, what are the process reengineering opportunities. That's how we have been able to look at opportunity and drive value.

So in most cases the clients come to us and say you are the domain expert, you know contracting all the way to care management, you also have the provider view, how are you able to bring this all together and that's a common question and this enterprise transformation is framework that we created is an answer to what our clients really wanted.

Even for larger clients if you just take the example that Roopam showed about the 6 million going to 20 million where the transformation saving is $5 million. And that is transformation becomes the key focus area on the managed services side. What are the different levers? We have built point solutions which are core which requires domain knowledge. I'll just take probably a couple of examples right.

The last one which says command centre for late payment interest. Today for one of the clients a claim needs to be paid based on local state guidelines within 30 days or within 25 days. If the claim does not get paid on time, then it incurs interest. For some of our clients the interest that they pay out of the claim, the leakage is close to around 50 million, 40 million, 30 million over the last three-four years. So we created an Appeals Command Centre. This is based on identifying pure opportunities based on our domain expertise.

Create process mining, create analytics framework to make sure that we know what are the bottlenecks, where are the claims getting stuck, why is it not getting processed, what are the challenges because of the platform

Page | 30

where the platform is not doing what it is supposed to do? It is not that the tech is not there, but it was not doing the work it was supposed to do. So, what needs to be done to trigger auto adjudication process higher?

We created that command centre, we looked at creating charters then brought in the automation led by Agentic AI to say hey these are areas where there are high bottlenecks, let's automate these which will reduce your exposure to these claims not getting processed on time which is thereby which you are reducing the late claim interest. For that particular use case for 2025, we saved close to around $15 million of late claim payment interest.

This is not an invoice reduction as Ramesh was explaining, not all value delivery is invoice reduction. This is the leakage that they were experiencing at an enterprise level. We have been able to reduce $15 million worth of saving based on the command centre on the late claim interest. Today we have taken that command centre on late claim interest to looking at appeals and say how do we solve for appeal?

Like that we have created multiple point solutions for some of the midmarket clients where their technology is not very robust and they have not matured and they don't have ability to change the platform in the managed services construct, these point solution becomes the value drivers for us to say hey if you are saving $5 million, where does that saving come from? It comes from these assets that we have created.

The layer below on the tech and AI led side, the technology team will spend a lot more time, so I'm not going to go into some of these areas that we are going to cover. But the other thing that I want to also call out is even before the Agentic AI what you call last one year of Agentic AI discussions that has been happening. Some of the value that we have given to our clients over the last four, five years has been very high. Say one of the Medicare national plans, our admin cost reduction has been close to around $11.5 million for last year.

Of this only 10% to 12% are probably invoice reduction. Most of it is enterprise transformation where we have looked at their overall administrative cost, internal other vendors and us put together and how have we been able to drive that admin cost reduction. In this case there was automation effort, there was what they called as a shift left strategy. A little technical but I'll explain a lot of the work that happens on payment integrity is as a result of things that could have been fixed upstream in the claims area. So, we looked at an opportunity of pure processor engineering.

Page | 31

It may not have be completely tech led or Gen AI. This was about looking at opportunity mapping the process using our domain expertise over the last 25 years we have been doing claims and we shifted all those and say hey what are the gaps that needs to be addressed, what are the changes that needs to be done in the platform and that project alone saved close to around $4.5 million which is part of this $11.4 million.

Another example is close to around $5 million benefit on this I talked about the example of the late claim interest. For a blue plan another cost avoidance project. So, this is something that we have been doing. The clients see that we have given these values and Roopam talked about our strong relationship and they have seen the domain expertise shine through with some of the projects that we have been able to do and that's how we have been able to go to our existing clients as well as prospective clients and say hey these are the use cases. This is how we drive the value. This is how we use the transformation levers in the managed services construct and make sure that we are able to pass on the savings to them. So that is how whatever the construct that you saw in the previous slide, how do we bring it to reality in terms of saving is all these levers and you will see the clinical side of it and the MLR side of it from Krithika's presentation.

Next I'll show one more demo. Next is about a demo and this is about HealthBridge Connect. I'll give a little bit of a context here. While we are looking at Gen AI close to 30%, 40% of our clients’ processes are very clunky still. They don't even have a digital platform. Forget Gen AI. So, we are also creating assets to say hey, what do we do for our clients in their journey of Gen AI how do we prep them.

There are some clients who say that hey I don't have portal where I can take bulk requests from providers to take care of their claims. That is a very big problem because my team is spending a lot of time taking these bulk requests. There is no portal. There are portals but it can't handle the volume at one single time. So now we created a portal which is HealthBridge Connect. Using this digital channel now we have made the process Agentic enabled.

Now we are able to use that intake channel, take all the requests from different providers, take the bulk request, have a different source then what happens with this is you will still have an Agentic process which will automate all these transactions but you still have also a digital portal which will make sure that whatever that fallout is which was earlier otherwise would become a call or it will otherwise become a claim you are able to manage that using a digital portal.

Page | 32

So that is the example that we are going to show in the next video. If you can go to the video about the HealthBridge Connect we can see what the power of that is.

Video Presentation Starts

I think one of the other questions that came about is, hey, how do we leverage what we do on the provider side? This entire workflow and the digital platform was created because we work on the provider side. We know what are the reasons and why they contact the provider, send those transactions, what are the fallouts, what doesn't get paid correctly on time and how do we make sure that we are able to mitigate that.

So, our payers appreciate the fact that we have this end-to-end view of the provider side as well. So if you look at the core admin side, 80% of the cost is driven by how the payer and the provider interact through all the different channels, claim submission, faxes, emails, chats, you name it right, all those and a single transaction from a payer and a provider across different channels on an average cost around $8 to $10.

By making alternate digital channels available and then fast forward then you make it agentic enabled means it's ready for agentic deployment you can reduce the cost by say close to 60%, 70%. And again, you are taking the enterprise volume and not just our wallet share of the volume that the client is giving and you are able to use these kind of point solutions in your managed services construct. So, when we say that, hey, we will be able to save 30% of your overall admin cost, these are the proof points and these are the solutions that we use to make sure that we are successful in those constructs.

Move from Managed Services to Synchrony. Managed Services point solutions the ability for you to impact in terms of the value is still dollars in terms of 15%, 20% incremental savings because you are still working on the client’s platform. You are still reliant on, they don't have the IT budget, they don't want to change the platform, you are still working, you are bringing AI on top of wherever you can to make sure and optimise the process and provide the saving.

In Synchrony you are bringing platform into play. I think Roopam talked about, this enables us to it becomes a force multiplier, right. Again, targeted to some of our clients and we acquired BroadPath last year. One of the things that we acquired as a capability was our member acquisition. Sagility historically was not in that space so we acquired member acquisition especially on the Medicare side.

Page | 33

Using that experience of what we acquired through BroadPath on the member acquisition, we looked at the lifecycle on the Medicare Advantage from member acquisition, enrolment and premium collections and billing. During this entire member enrolment cycle, which is very high during the peak season from November to say January, February where you have a lot of volume of work that increases. How are you able to help your mid-market clients to bring this all together?

Today they go to different partners, they have some amount of leverage in terms of bringing these technology solutions and some of these processes are still manual. There is a lot of regulatory requirements also. You have to make sure that you are submitting all these records once a member's plan is built to make sure that the enrolment kicks in. You want to make sure that you test the benefits that have been brought in are tested so that the member's benefit is correctly reflected for the claim’s payment.

We bring all these together some of the partner solution build bring in the AI layer on top of it to make sure that for a client when we say, hey, we will do the synchrony the cost savings is much more than what they would have gone with point solutions. That's one example of how we have used on the Medicare Advantage side a Synchrony solution. Here we have partnered with two, one on the member enrolment side where they bring their enrolment platform Benefit one and then on the billing and collections area and then our own member acquisition which is on the services side that we have always been doing using the BroadPath experience.

Today this is something that we have launched as a productised offering and we are ready for next open enrolment which is the next season coming up in 2026. We are already working on a go-to-market plan in terms of, hey, what are the different Medicare Advantage plan in our existing portfolio that we have, where we just or not just focusing on the member acquisition space. Historically that's what we did. Now we can offer an end-to-end solution through the Synchrony solution.

So one, faster value for our clients in terms of immediate savings. We are able to bring this all together and we have for the client Sagility orchestrates this entire solution together and also stickiness where we are not just looking at services, but we are also bringing the platform which increases the stickiness where they will not be able to change the platform for a short-term basis.

Page | 34

So that's one of the examples of the Synchrony solutions. Similarly, Roopam talked about Synchrony using capability on the claim side. You are looking at Synchrony for the appeals and grievances process. So this is how whatever capabilities we have, we bring the agentic layer, orchestrate it, use our domain expertise to say, hey, what are the opportunities where things are not working, how do you reengineer the process, bring the technology platforms and thereby commit to the savings that we are providing to our clients.

So, two examples, one on the managed services side, one on the synchrony side. I will pause couple of questions if at all. I don't mind, thanks.

• Participant:

• Srikanth L:

• Participant:

Srikanth L:

Are we going to the path which PBM had gone 10, 15, 20 years back?

This is not necessarily PBM per se. This is even today we structure plans whenever a plan is getting sold if it is mid-market or if it is Medicare Advantage. On the services side we construct plans, test it out etc. But we did not have a platform which will take it and say, hey, where do you build a plan on and how will the benefit structure look like based on the enrolment. That is what we are bringing with the platform. The services was something that we were already doing. So now we are looking at the platform to bring in and looking at both the services as well as the technology platform together.

In previous slide you mentioned about the cost saving effort which you showed to your client. So just want to understand the cost saving is mainly a function of elimination of repetitive task from the client ecosystem and showing them the benefit or it's structural challenges which they are facing in the operation?

• It's a combination of both. Structural challenges, if you look at all these costs, most of it is structural challenges. Repetitive task is the admin portion where the invoice reduction I said were out of this 11 million probably some of the repeated tasks where we are automating and ensuring that it gets automated that will be the repetitive task portion. But most of it will be at an enterprise level where you are looking at leakages that they were having because some of the volume that was being driven by either the platform not doing what it is supposed to do with a claim engine or if the process was broken, because you had an enrolment team looking at a different set of SOP versus a claims team looking at a different set of SOP which were not talking to each other, so we bring in our domain expertise.

Page | 35

So those would be the larger ones which contribute to these savings rather than just the repetitive task of RPA. That number in the past used to be higher, but since we ourselves have automated quite a bit of these processes, those repetitive tasks using RPA etc have reduced. Now it has become a lot larger in terms of how we are really looking at cost avoidance for our clients that they are incurring.

Participant:

Let's say client rollout AI internally to find the leakage in their process, then what going to be impact on this?

Srikanth L:

Today we are doing that for some of our clients where we have identified leakages for the clients based on how as I was explaining where we were the innovation leaders for one of the clients. That's exactly the exercise where they use us to say, hey, this is you know our process better. You tell us where are the leakages. Wherever there are AI infused technology that we can bring in, we say, hey, a claim assist solution requires AI, this is the agentic framework, let's deploy. Then we go to the next level of saying from a pure identifying that opportunity go into deploying those claim assist solution using agentic AI.

But if the client has already identified a project and identified an AI led solution that they want to deploy, they still come to us and we become a subject matter experts in terms of helping them identify which are the areas where the time to value is much more faster because they don't know the process as much as we know. So, they come to us and say, hey, what are the different levers? Is this a particular candidate which will give you 5% saving, 20% saving? Those are conversations we have as well if at all they have their in-house projects that they are driving. They use us for those subject matter expertise.

Participant: One question, what would be the share of managed services and Synchrony to the revenue today?

Srikanth L:

Maybe that can be answered in the Q&A session.

Participant:

So we talk of these numbers, 11.45 and 4.5. If you can give us a reference to what it means to the overall contract size or the overall savings and what is the scalable up to that probably give a better idea what the extent of the benefit is actually.

Srikanth L:

I'll give an example, that's why it's a little tricky also. Not all these $11.4 million saving comes from our contract with them. If we might be doing a work of say $25 million worth of book of business. In that 11.4 million does not mean that all of that was our value because we deployed all these solutions at an enterprise level. That's the main premise of the managed

Page | 36

services. The entire transformation is at an enterprise level, so it is our volume, it is the client's own volume, it is some other vendor's volume. So that's how the construct is. It is not just about arc, so generally whatever we say if on the managed services side if at all you are looking at the example that Roopam talked about, we have committed to a saving of around 5 million over 20 million that's typically a percentage that we look at, because some of these are point solutions that we bring, whereas in Synchrony it's a lot more higher.

Okay, I think we break for tea or coffee. Thank you.

Dhaivat Mehta:

We'll come back to the hall by 4:25 p.m. so that we can start the session on time. Thank you.

[Tea Break]

Dhaivat Mehta:

Hey, everyone. I hope everyone's had a good cup of coffee and tea. We want to make sure that we stick to our scheduled time. So I request everyone in the room to please take their seats. Gentlemen at the back, if I can have your attention, please. Gentlemen at the back. And I request my colleagues outside to please help usher our delegates in the room.

Gentlemen at the back. Hi, yes, thank you. We have a wonderful session coming up with Krithika who will talk all about Clinical. And Krithika you have the pleasure of the after-tea session. So, no pressure there. Can I request our friends to please have a seat? And CDR team, can we also request them to please help usher everyone inside? We'll have lots of chance for more conversations over dinner. Okay, Ashwini, most of our delegates are inside. Okay.

Krithika, over to you.

Krithika Srivats:

Good afternoon. Can everyone hear me? They can’t hear me. They can? Now you can hear me? In the back? Excellent! Good afternoon! My name is Krithika Srivats, I lead the Clinical Practice. My colleague Sri told me that I can't be too technical and so I asked AI to give me a version that's not as technical.

It came back and said, don't worry, your English is at fifth grade level. So take it as hopefully it's not filled with jargon. Anyway, Sri spoke about payer-provider synergy. Keeping the payer and provider issues in the centre and how can we bring enterprise transformation. My story today, what I'm here to talk about is keeping the patient in the centre of it.

Page | 37

Roopam talked about MLR rising. There's a lot of consumerism, aging population, a lot of things that's happening in the industry that's driving up the medical spend. So we're going to try and take an attempt at what that patient experience is and why some of these patient experiences that are not unique are actually rising the medical cost and shifting the MLR. So with that, we're going to talk today a little bit about a story of Graying Line.

Meet Abigail DeYoung. She's a 67 year old lady. She lives alone. She lives in a two story apartment. She has some financial insecurities. As a result she doesn't have transportation, she doesn't have a whole lot of support. She's at 67, still dependent to bring her own bread home. She has diabetes, she has hypertension and off recent she has been experiencing a lot of back pain. And anybody who has had back pain would know it's not something that just comes and disappears. So, like everybody else she has tried to put up with it, she has tried her conservative measures, but nothing has worked and now she is really in need of, some intervention because she's no longer able to move and perform her daily activities.

So, if you look at the problems that Abigail has in numbers, Abigail is not alone. 75% of medical spend in the U.S. comes from people who have multiple comorbidities, at least two comorbidities. There are several access delays when there is social determinants of health. That's when people have not sufficient access to care, not sufficient understanding of care needs as well as other financial and educational and other type of constraint.

And they could cost anywhere from – intervening for back pain could cost anywhere from $45,000 to $70,000, $75,000 and at an average Medicare members cost about $1.2 trillion in the U.S. So the opportunity to really impact is pretty huge. But it's not just that, if you look at medical spend it doesn't come just from healthcare issues.

I also mentioned about access to care. When you have a clunky system that doesn't make it seamless to access care that definitely drives up the medical cost. So they say that some of you talked about provider using generative AI processes for providers. While that is happening in the industry as it stands, 100,000 nursing hours just to get prior authorisations. And these numbers here by in itself may or may not make too much sense but when you look at the big larger context of managing 75 million Abigails, you can do the math, 60% or so of services are still very manually transacted between the payers and providers.

Page | 38

Very rarely does anybody get complete information to make decisions. So when we talk about generative AI opportunities, they are huge in bringing together the longitudinal view of members today.

So all of this clunkiness not only result in; say for example, prior authorisations typically take 7 to 14 days to get approval before you go get the care but then by the time you add the number of prior authorisations, she's going to need whether it's for diagnostics, whether it's for conservative treatment like physical therapy, pain management, and then finally if she ends up needing surgery – you can multiply how many 7 times or 14 day times these services that she's going to have to wait for, really delaying her care to like three, four months.

So how do some of these solutions that we have, you heard Roopam and Sri talk about Synchrony Solutions, SmarTec solutions. How do we enable our services, our clinicians who are performing these prior authorisations and care management services. How do you enable them to be more efficient but also be more effective? Some of how we're going to talk through to understand what Abi's experience would look like and what type of impact we could create.

So human expertise meets algorithmic precision and that is the reason we titled it this way because, everybody, if Abigail went to a service provider being managed by Sagility or anybody else, the process remains the same. The provider has to request a prior authorisation, they have to review if it's medically necessary and they have to then perform the necessary evaluation and then make a determination. What we are trying to say is there are two ways how you can impact medical costs.

One is by addressing the waste and that's called utilisation management. You look at unnecessary care, but unnecessary care is just by itself is subjective. For example, if I come from a background where I'm used to self-medicating, I'm used to dealing with pain, my pain tolerance is very high, I may wait for a long time to get care. I may not need a whole bunch of other; I may not try physical therapy, I may not try pain injections, I may directly wait for so long that finally what is needed is a surgery.

But if I'm also somebody who is very astute about how to care for myself, I would constantly be going to the doctor even if it's slightest pain I'm going to be constantly going to the doctor. So waste is not just, there's not a universal definition. It's very subjective, that's why it's very important to meet the patients or the members where they are. In the context of how their lifestyle is, in the context of what they understand of healthcare, of

Page | 39

their own health, how to access health, as well as in the context of some of their social and behavioural issues.

Ramesh spoke about social determinants of health and social and behavioural issues in his slide and the context of exactly this, if I have all it takes, if I have a caregiver at home, I'm not the breadwinner, the options of how I would go about my surgery is very different than when I'm the sole breadwinner and I'm the only person caring for myself.

So what we do today at Sagility is we manage the waste part of it through an integrated program called utilisation management. But we also have a quality of care solution, which I'm going to be talking about in the next slide, which addresses not a one and done. Nobody is one and done when you have chronic conditions, when you have extensive healthcare issues. It's not a one and done. So how do you simplify getting the member what they need through a program that's not one and done, then it's not too repetitive and it's not unnecessarily adding more cost to the system.

So in this case the algorithmic precision comes from having a mechanism to identify Abigail in the context of who she is, not just that she has low back pain and not just that in this case her provider has recommended her for her to have a surgery. So what happens now is Abigail goes to the doctor, let's assume that she's had some of the conservative measures, where she is right now, her primary care doctor says you really need an ortho surgeon and the ortho surgeon said you need two, three levels of surgery.

Your back pain is because of a lot of issues in your spine, there's three levels for which I need to do a fusion. So the request comes in, typically requests comes in either via contact centre or via a Fax based solution. There are provider portals, she talked about it.

Some of them can handle a lot of information, data exchange, more digitised format and others are not. We have a partner who has about a 90 plus percent adoption of portal. So right off the get go, that's a win solution because in the back of that portal we are able to enable that with algorithmic decision making. So at Sagility, we have a more digital front door approach, whether it's coming through a Fax or through a portal. And we're going to be seeing some of the demos for Ram and I are going to be talking a little bit about certain solutions which actually extracts all the information from facts and enables digitisation of the request in itself.

The reason for that is two parts. One is it significantly increases the decision timelines. I said 7 to 14 days earlier. There is a CMS regulation

Page | 40

that says that everybody has to make transparent and timely decisions still within 7 days. That's a significant improvement from where we are today but imagine waiting 7 days just to get a diagnosis done for any of these procedures.

But this enables us to at least meet the demands right now. Once the provider submits the request, there is something called medical policies. These are evidence based guidelines and what that does is it enables clinicians to have a standardised protocol to review and see if the service is actually needed. If the service is needed, you have to go through about 20, 30 pages of guidelines and they are not in the form of questions or anything. So the nurses have to interpret and make accurate decisions based on reviewing those guidelines.

What we've done is simplified that into an algorithm and so that when the provider is submitting the request, they answer three to five questions. And in 70% of the cases, they're able to get a decision in real time. If they don't meet the criteria, then it gets routed to a nurse. And the nurse again uses some generative AI solutions to streamline the process of review and we're going to demo that. And if it doesn't still meet the criteria, it actually goes to a MD who picks up the phone and he's talking to the surgeon and they're discussing to see does she really need the Lumbarexctomy?

And if she does, does she really need three levels of Lumbarexctomy? Let's contextualise that based on where Abigail is and then they make that decision. But what happens along the way is because the data has been digitised right from the beginning, all of this is summarised with which criteria was met, which was failed. So that when the provider is calling, the MD is calling the provider, they're looking at transparent decision making process.

And so now you're actually reducing the payer provider abrasion, the friction points, because you're collaborative, not because you picked up the phone and called, but because the criteria that you're evaluating that patient against are fairly transparent.

So this process allowed us to do two things. It allowed us to actually look at the patient from the lens of not just a lumbar surgery. It actually allowed us to look at the patient in the context of what their social needs are, how they're going to cope with their surgery, because that's just as critical because I think Sri or Roopam mentioned about readmission.

Page | 41

That is a very big, there's about 30% of hospital admissions for people with chronic conditions that causes readmission within 30 days. So that's large, both in terms of spend in itself and the management of that process is multiple rework. So through this process, what we were able to then determine is that this member does not have the ability to handle a multilevel surgery because she's the sole breadwinner. You want her on her feet much faster. It also gives you the opportunity to see if she's going to continuously stay with her protocol.

So for this patient, one level surgery is more than enough. So not only have you reduced the cost, but you've actually kept the patient in the centre of that decision making process. And you will see in my next slide how this actually converts into not just reducing unit cost at the line item level here, but you're also going to see how that actually reduces further readmission cost and prevents future admission cost.

Couple of things I want to call out here, our partnerships as well as some of our point solutions enable us to have a very robust value addition to the mid-market. So you see on the right bottom here, the availability portal is a solution that can really drastically reduce manual transactions between payer and the provider. And that is the higher the volume, the better the economies from that particular solution. So it's a large enterprise sort of model today.

But we also have partnerships for the mid-market. Our mid-market solutions really aim at a couple of things. One is in clinical, unlike a lot of the administrative process, every clinical program needs a fairly high cost domain expertise. So for example, if you're managing oncology, you need an oncology team. You're managing musculoskeletal symptoms, neurologies, whatever you have to have specialties, specialists from each of those areas.

But when we are bringing multi-domain and we are simplifying those programs, you need to have the ability to manage it through very niche, targeted, multi-payer sort of an environment. So some of our partners actually provide that capability for us. So with this, in this particular example, Abigail has received her decision to have the surgery within two to three days – significantly improves provider experience, reduces friction with the provider, improves member experience.

Any of you that are familiar with star rating, there is a direct correlation from a survey for member experience. And it's not coming from managing their interactions, it's coming from, I mean it's not coming from handling

Page | 42

their calls, it's coming from managing interactions across the value stream.

So UM typically can reduce anywhere between 5% to 15% of medical spend. So something to just sort of keep in mind, a good integrated UM program and the opportunity in the industry is about $200 billion to $300 billion.

Now let's look at how Abigail goes through bending the cost curve through the quality of care programs. Quality of care programs, the intent is a couple of things. One is clinical outcome, that's the most important. But Medicare plans especially, as well as across all lines of business, there is some of you may have heard of this term called HEDIS.

Essentially what it is healthcare effectiveness data being coded and managed proactively in a preventive manner. And the more you actually identify and manage preventive health, you get better ratings than your star rating, your quality bonus. So in this case, we know we identified Abigail to have diabetes and hypertension, not because of some risk analytics, but because we got one interaction through a prior authorisation medical record.

So typically what happens, today as it is, if you talk to plan, they manage, they would have approved or denied that lumbar surgery based on whether it met the medical necessity criteria, so on and so forth. But that would have been a one and done close transaction.

Then somebody is chasing Abigail because their risk identification flagged her because she has two comorbidities, diabetes and hypertension. So now they are calling her 20 times because they need her to go to the doctor and manage her diabetes, get her A1Cs tested, control her hypertension. So there is a whole another program that's working just making outreaches to her. Then on top of that, you saw our integrated UM in the prior slide that got her to at least get the right level of care. But then post discharge is managed by another team typically. They follow-up and say, hey, you went to the hospital, you had a surgery, hope you have everything.

Here is Abigail, who's just come back, she's trying to manage herself and she's had a pretty major surgery. Imagine getting called 20 times just to say, “Hey, do you have everything?” Why can't we embed that into an integrated UM program so that we flag right away saying that she's going to need help when she goes home. Instead of sending her to another

Page | 43

rehab facility, delaying care, we know she has to get back on her feet very quickly.

Can we send her home with home health? Can we proactively recommend some physical therapy that's going to get her back on her feet? So, a good rounded post discharge care management program shouldn't be a standalone activity. It shouldn't come because somebody did some analytics, they got a trigger saying, so and so was in hospital. It should be seamlessly integrated into that care meeting, the member where they are.

And then last but not the least, we also have in Roopam's slide, you saw our Synchrony Aging in Place program. Today, aging in place is largely in the preventing long-term care admissions to facility-based care. But in a post-acute care, what post-acute care is essentially, acute care is anything that is right there done in the hospital. And then when they are discharged from the hospital, everything that follows, that's post-acute care. So by identifying other needs that's going to keep her at home safely, whether now she has fall risk maybe or she has two story home with her back pain if she's going up and down without the right type of training or equipment, she's going to fall and come right back to the hospital.

It may be a different episode, but nonetheless she's going to be back at the hospital. So looking at what other type of needs she has so that she can continue to stay functionally independent at home is our long-term care, long-term resilience program. So an integrated solution that is enabled by AI, predictive models.

Today predictive models are more intelligent. The way we are looking at it is using Generative AI to mine information that comes across all interaction types. Whether that's coming because of a prior authorisation, it's coming in because there was a call that was placed specifically to ask for where this provider for back surgery would be available. So sort of unstructured data across all channels, you have the ability to pull all that data together and then the domain expertise comes in really knowing how to develop the ontology, how to create those queries and risk stratification that can then drive each of these care pathways.

So the result of program like this clearly start rating, which results in very huge quality bonus payment, lower utilisation and I said about 5% to 15% impact on the medical spend, 30% lowered admission rates and more importantly programs like this really have 3x to 5x ROI and I'm going to talk

Page | 44

about that in just a second. So this is what is called as whole person health.

You saw in Ramesh's slide he talked about whole person care. Understanding root cause of poor health and managing it effectively through a combination of tools and technology available through better predictions, better predictive models and then having a workflow that can be seamlessly integrated between all these siloed functions, because today these are all independently managed siloed functions.

This is an example of a recent engagement with the client. There are a lot of numbers, I won't bore you with all of that. The very high level story is the client was spending, the client has elderly population with a lot of special needs. And their spend was about 3x, 3.5x more. Their medical spend was 3x, 3.5x more than their peers in the market and they were in sanction.

What we looked at was we went and said, what are the opportunities to control medical spend? Out of the 3x increases, we identified two or three areas and one of it was home health. If you have somebody who is receiving home health services, the expectation is that your overall total cost of care is lesser than somebody who is not receiving home health care services. Yeah?

In this case, we went on to see where is; when people are getting home health care services, first of all, are they costing you lesser than somebody who is not getting home health service in the similar cohort? And if yes, what is that additional opportunity? If no, what should they be doing? We looked at it under two vectors. Vector one is called frailty risk and vector two is if they have behavioural health issues like dementia or depression, so on and so forth.

People on the left side who are lesser on the frailty risk, utilisation management is a good program that can control costs. Except you see here people who have no behavioural health issues, lesser on the medical and frailty risk, still constitutes to 3.6x more cost compared to somebody exactly like them not getting home health care service, which means your UM is not working. So somebody was asking about competition and how are your programs different and are you able to add value. Right there, this is a program that's done by competition. So you still identify about 8, 10 million opportunity when you bring in the vectors of predictive analytics, when you bring in that patient centric model.

On the top, that's about okay, it's comparable. You're receiving home health, you're not too far off from somebody who's not receiving home

Page | 45

health. So your UM, you can say, is working or at least it's not working. But on the right side, there are a lot of issues. On the right side, while the home health, the people who are receiving services may or may not be doing better or worse than the other, these are people if you don't manage them today with just good UM or an integrated care and case management, they are going to be coming back as future spend for us.

So really bundling care management as well as what this tells you is if you have somebody spoke about a $6 million UM operations. This is exactly that. This client was spending about $6 million in UM operation. But if their UM wasn't working, can we actually target and create cohorts and do the same program differently, so that people who need to be in the respective buckets are getting the right type of care? So you're rightsizing your operations.

So this is a bundle on the right side. We actually bundle some of these solutions with care management. On the left side, we just enabled and enhanced their UM integrated program, resulting in about a $24 million savings for not spending too much more on the UM operation.

So with that, I'm going to just start a video of one of our solutions, a generative AI solution. Ramesh mentioned about our award-winning Nurse Assist. So we recently got named in the Intelligence Excellence award by business intelligence group for our Nurse Assist. So you're going to see a video. The video just sort of tells you what the construct of the Nurse Assist is, but then we're actually going to be doing a demo in Ram's session. So with that, can you please turn the video on?

Video Presentation Starts

Thank you. Madan over to you. I think I'm right at time, thirty seconds to go. So, we'll park the questions for later. Thank you.

Madan Moudgal:

Can everyone hear me? Okay. Good afternoon, good evening everyone. My name is Madan Moudgal, and I along with my colleague Ram Natarajan are going to take you through a journey of our technology capabilities. And you heard by now about the fact that our technology capabilities is one of the pillars of what makes us different and will ultimately ensure our success into the future.

You also heard quite a bit about the fact that we claim a lot of credit for our domain knowledge. And that extends by the way into how we deliver our technology as well. So even when it comes to technology solutions, technology development, we rely heavily on our colleagues on the domain side. People like Krithika that you just heard from Sri and so many

Page | 46

other people within the organisation who allow us the benefit of their knowledge to make sure that we're able to deliver very cutting-edge technology that is suited to our particular industry.

Okay. So let me talk a little bit about our investment strategy into technology. Over the course of the last 10 years, we have built up what we believe is a fairly robust set of capabilities and we refer to that as our technology portfolio. Now these capabilities, think of them as building blocks. Now these are the building blocks that allow us to very rapidly assemble point solutions, very rapidly assemble solutions to address a particular business problem that comes our way, right.

And by the way, a lot of these technologies have been built up over the course of time. And by taking advantage of technology that's been available at that point in time. So think of it as a cumulative effect of having all of these technologies now available to us, as we now take on the next opportunity, which is brought to us obviously by Gen AI and Agentic AI. But Agentic AI by the way is not necessarily the solution for everything, for all of the problems that are out there. And I'm going to make a point about that as we move forward.

It certainly is a very, very powerful tool and we intend to use it as such. But clearly, we are also going to be taking advantage of some of the other key assets that we have within our portfolio. One other point I want to make is that, in the delivery of some of our capabilities and Roopam talked about Managed Services, we talked about our Synchrony solutions and so on. There is reference to enterprise applications as well.

Now we are not in the business of building enterprise applications and by definition enterprise application in our world could be like a claims platform. It could be a membership system, it could be a CRM. We don't build those. Rather, we partner with vendors who actually build those out. And where necessary, we integrate our capabilities with theirs to deliver those end-to-end solutions that Roopam, Sri, Krithika and others talked about.

So just in terms of our overall technology strategy, I wanted to make that point that we build componentry, if you will, modular building blocks. And those then come together are assembled by us, leveraging our domain knowledge to deliver cutting edge custom, if you will, in a lot of cases solutions. But at the same time, using components that are reusable, so that we don't have to start from scratch every time.

Page | 47

Okay, having given you that kind of an outline, I want to be a little bit more specific. So let's talk about what I mean by these components. And again remember I said earlier that this has been tools and capabilities that have been developed over the course of 10 plus years. Yeah. So it includes things like data models. I know Krithika just spoke about it, Sri earlier spoke about predictive models. So predictive AI has been out there for a number of years now and we have taken advantage of predictive AI to build some very cutting-edge algorithms and models and very effectively use them. And we intend to continue to use them into the future, because there are some problem statements for which a statistically based predictive model is more appropriate and a better solution than a Gen AI model.

And that's something to keep in mind. Like I said Gen AI is not necessarily the answer to everything. It certainly is the answer to a number of our problem statements that we have. We also have developed point solutions. And I'll talk a little bit more about some of the examples here. We have been leveraging RPA, Robotic Process Automation. That's been out there for a decade or more. And we have certainly taken full advantage of it to build a library of automation bots. And by the way, because of the fact that we are dealing with legacy applications in our industry, large in some case still mainframe run applications, a process automation bot is probably a better answer in some cases to automate a particular function.

So that's why we always look at where we need to apply our capabilities and then decide what technology makes sense. Now we certainly have not just delved into, but heavily invested in the Gen AI space. Over the last 24 months, we even made an acquisition, which you all probably know about in the generative AI space, brought some capabilities in, incorporated into everything else that we've been doing for a while now. And consequently, we have both in what we describe as our front office as well as in our back office, we have built a number of different solution sets for automation using generative AI capabilities, including large language models, small language models, and we'll give you a few examples of that.

Dashboards. Sri spoke earlier about LPI – late payment interest, right. So there are some capabilities that we had to put together in order to solve that particular problem. But ultimately, we need to deliver all of this information in some cogent, in some organised way. So that people can then take action on it. That's where dashboards become very important. And these dashboards by the way, when we build them, we again build

Page | 48

them with the intent that we'd like to apply these core capabilities to different types of use cases.

It may be in the one case solving an LPI problem, a late payment interest problem. In other case, it could be around managing appeals more effectively. So that's where the advantage of building things from the standpoint of configurability, reusability becomes so important. When we go about building something, we ideally would not like to start from a blank sheet of paper, because that takes the longest. So what we look to do is, we look to accelerators, we look to frameworks that if you in a sense provide you the skeletal structure on which you can assemble various capabilities.

So to that end, we have also invested in building out a few frameworks and accelerators. Now granted, they are not necessarily all homegrown, in some cases, we are also experimenting with accelerators that vendor partners have made available to us. And we constantly are on the lookout for best of breed capabilities there. Interoperability is a very big opportunity and a challenge within our industry. We talked earlier, some other speakers referred to disparate systems, legacy applications, the complexity of payer and providers interacting with each other and so on and so forth.

There was a question earlier about I think from someone here about the fact that providers are leveraging Gen AI quite extensively in order to level the playing field with payers. And by taking advantage of Gen AI, they are now automating some of the things that they do. For sure they are. I will give you a simple example. When they are reaching out to a payer to get a question addressed. Now they are making robotic calls rather than having to pick up the phone and call because that's more labour intensive. So payers get inundated with lots and lots of calls coming in and that becomes a bit of a challenge for them. There are solutions to that problem as well.

But if you think about it, a solution could be that you respond to a robotic call with another robot answering the call. But that's kind of a very clunky way of solving your problem. A simpler solution could be that if you had an API that the provider could reach into to get their problem answered through an API request and the API response with your answer. In a sense, HealthBridge Connect is getting us towards solutions that are a little bit more elegant, right? So we need to think in terms of the capabilities that are available to us, apply our domain knowledge and then come up with these solutions that make the most sense.

Page | 49

There is another element too, to keep in mind, which is cost. These solutions don't necessarily are not free. Particularly if you look at Gen AI, if you were to attempt to apply Gen AI for every last problem and you were to build it on the backs of these foundational models, especially the large language models, the cost can run up. So you have to be more careful in terms of how you select the right models, the ones that are most, that balance cost with accuracy, with latency. So these are all criteria that we look into, that we take the time to research and figure out what makes sense.

And that's where some of these frameworks come in, where you have the communication manager. The framework there allows us to switch out different LLMs, switch out different SLMs, small language models, where it makes sense. Now, we also have this reference to what we are calling clinical language models, where we look at models that are more adept at understanding medical terminology. So these are all the types of capabilities that we look into to make sure that ultimately the types of solutions that Krithika is looking to build within her practice area or that Sri is looking to build that it is working well. And that we can deploy it and also deploy it in such a manner that our clients are comfortable that we are securing their data, that we are securing their interest in a sense.

And now to that end, one of the things that we have made efforts to do is to ensure that our engineering, our software engineering efforts are robust. That they follow very specific policies and procedures and that the controls are in place, so that we are securing our clients’ assets. And what we have done in that regard is to ensure that we are SOC2 compliant and HiTrust compliant for all this work that we do.

Now, for those of you that are aware of this, this is in a sense a gold standard and a certification that gives a level of comfort to our customers that we have their best interest when it comes to managing their data and delivering these types of solutions to them.

So I talked already about a few of these examples and the types of things that we're doing. So I'm not going to belabour this point, because I know it was also referred to earlier in some of the earlier speakers' comments. But this should give you a sense of the places in which we make our investments. And we will continue to make these investments as we move forward. Agentic AI is now very much in the fore and so we are starting to make further investments in Agentic AI and I am going to talk about an example of how that is coming together in order to deliver on a specific use case, okay.

Page | 50

Before I go there though, I wanted to also talk about the other point, which is our partner ecosystem, because if you recall, at my opening slide, I talked about the fact that when it comes to enterprise applications, we don't necessarily develop them, we work in tandem with our partners. So here, to give you a sense of our partner strategy, there is some platform vendors and you see some names up here, our platform vendors that we work with. And of course, as we deliver these managed services type place, there is a need for us to, and the Synchrony solutions, a need for us to implement some of these systems and to integrate to them. So we work with; we don't necessarily have at all times maintain a bench strength to be able to scale up all of these. It doesn't make sense in our business.

So what we do is we have a partner a robust partnership strategy where we partner with systems integration partners like the ones you see up here. There are other capability partners. I referred to robotic process automation. We've been doing it for a while and you all probably know that Automation Anywhere is a very well-known name in that space. And we've been working with their tool set for quite some time now. And like that there are some other vendor platforms as well that we're very adept at using. And of course I cannot not mention the hyperscalers. AWS and Azure are certainly two of the largest hyperscalers that in our industry, there's a heavy a lot of comfort with amongst our clients.

So we have ensured that our solutions the ones that I was talking about earlier are deployable and maintainable within these platforms and that in a lot of cases we are also taking advantage of the tech stack that these platforms provide us.

Okay. I'm going to now double click a little bit just on Gen AI and Agentic AI for the obvious reasons, right. So one of the things that we have focused on in our strategy for how we are using Gen AI is to think of it in the context of a synergy between the agent, the virtual agent and the human agent. And I will use that term just to keep the two separate.

So there is a human agent and then there is a virtual agent. And we believe very firmly that in the future, they need to work in Synchrony. And so to make that possible, there has to be, if you will, a system of engagement between the two. So that you can coordinate and in a seamless fashion move information and task and triggers between the two.

And so there is, in a sense an interaction platform that enables that. And that's starting to be borne out in the types of examples you see below of some of the use cases where this synchronization is working very

Page | 51

effectively. And we continue to add like we have done in the case of the process automation, the robotic process automation arena. Now even in the Gen AI arena, we're starting to add to the number of use cases of application of Generative AI.

So whether you're talking about UM intake, the intake process that Krithika was referring to or you're talking about, let us say we are talking about appeal status and where someone is calling in to understand what is the status of an appeal that they have submitted. That is something that today is purely a manual process. But tomorrow it's likely going to be not just a human agent involved, is very much likely going to be a virtual agent involved as well. So these are all different examples of how we're applying the system of engagement model to enable this combination of working together.

Okay, I'm going to conclude by taking you through something a little bit more detailed in terms of how we look at Agentic AI. And it's through the use of an example. Yeah. So prior authorisation, utilisation management.

In the utilisation management world, a provider submits a prior authorisation request, prior to the service having been performed, asking for authorisation to move ahead with that service. So the prior documentation is submitted. And today that prior authorisation request, it can come through multiple channels. We still use Fax, believe it or not, in our industry, quite a bit of it in fact. And so it can come through Fax, it can come through paper, email, what have you, right?

Now what we have envisioned and what we are building is an orchestrator bot that is going to orchestrate a sequence of events that are going to take place. Just like you have today a human agent that receives this information and starts off a process. So there is first a need to do something called case creation, which is where you initiate this process by the creation of a case that again gets followed through to its logical conclusion. When this document comes in this prior document, it's got certain details on it.

Those details need to be extracted. You can do it the old fashioned way, which is you can read it, you can then key from image, if you will, into another system or you can extract that information. You see reference to OCR and NLP. Optical Character Recognition, Natural Language Processing. These are technologies that have been around for quite some time.

Page | 52

We can still take advantage of those. But we can; the entity that's managing that OCR process is not a human anymore. It's a Gen AI bot. In this case, a case extractor bot, which then triggers off a process after the information has been extracted to a case generator bot that summarises that information. Because it needs to be summarised as part of our process and then initiates within a case management system the creation of a case.

And there is a process for eligibility verification that can be handed off to another bot. But if you see at the end of this, you don't see a bot, but you see a human. Because according to this process, it still requires a human to authorise this particular case. Now the human may decide that this warrants some further documentation to be received, in the form of let's say, they demand medical records before they provide approval for this prior authorisation.

So there is an entire medical records management process that kicks off. Where you have a medical records manager, another bot. And that then results in the submission of a voluminous document sometimes it could be hundreds of pages long. And now you have to extract information from it.

You have to summarise it. You have to annotate it. So that it is easy for the nurse practitioner who is at the end of this process to make sense of it. Today that nurse practitioner probably spends 30 minutes pouring through this detailed document, hunting and pecking and looking for relevant information. No longer necessary because it's already been done for you.

So the bot's taking care of that for you. And then there's the key step of the final decisioning. Even here, Gen AI can certainly play a role in referencing clinical guidelines, which are the rules that you consult in order to decide what to do with this particular case. And you could have the bot make a recommendation. But it's only a recommendation.

The final decisioning on that recommendation is left up to another human. So you see the symbiotic way in which this is all being done. So this is an example of how we're orchestrating using Agentic AI an entire end-to-end process, which clearly shows bots and humans working together.

Okay. What we are going to do now is I'm going to turn it over to my colleague Ram who is actually going to take us through a little bit more of a deep dive into this through by use of an example, right. But before I turn

Page | 53

it over to him, this was my last point. I have about three minutes. I could probably take one or two questions. No, no questions? According to this I have three minutes. Okay.

Ram Mohan Natarajan:

Can you hear me now? Thank you. Thanks, Madan. Good evening. My name is Ram Mohan. And I'm going to spend the next 15, 20 minutes with some of the agents that I'll show on screen on what we've been talking about so far. So we'll just take the same example that Madan spoke about, which is about prior authorisation. Take a scenario where the prior authorisation has got denied. So what happens is that the hospital is now going to send additional documents.

They're going to get additional proof. They're going to send medical records and all that, and that's what is called an appeal. Like, Sri spoke about the cost of a claim being about $10 to $20. But if you see some numbers here, the costs start multiplying significantly when it comes back as an appeal, because now it's going to come with much more complicated cases. And this could be one example, like I said, was on prior authorisation. It could be about a claim having 1,000 lines, and two of the lines were not paid correctly, and an appeal can come for that. It could be that there is something that the insurance saying that it is not medically necessary to do a surgery.

So there could be multiple reasons why this could come. And like you can see, it's an important part because the costs start increasing significantly. And the reason it's relevant is also because a lot of times when you think about admin costs, we're talking about one person sitting and doing a process costing X amount of money, whereas what we were talking about earlier on the clinical side and others, the impact of all this is significant. If you don't do it right, you're talking about some huge legal issue. You're talking about compliance issue, all running to multimillions of dollars.

So it's important to keep that context in mind. The other thing which is also important to keep in mind is that while technically you can use AI to do a lot of stuff, there are a lot of stuff that are also governed by compliance. We need to have explainability. We need to have the right documentation. And we'll show you how we have used AI and humans working together to address this problem.

So I won't get into too much more details about Star. I'll try and touch on it, if time permits, when we go through a demo. What I wanted to show is a quick video to set some more context on what the appeals is, and then we'll talk about various agents performing this job.

Page | 54

Video Presentation Starts

All right. So, what you saw were actually screenshots of typically how an appeal document comes. You might have appeal documents like you saw in that example, over 100 pages. You will have handwritten notes. You will have documents with the doctor's signatures and the patient agreeing on waiver of liability. So there's a whole lot of attachments and documents that come there. So what we've got here are a set of agents who can these are all AI agents which can do the work. I'll just talk about two, three examples to just give you a flavour of how this whole Agentic system works. So the first one, Madan spoke about it too, is the intake agent.

So what does this agent do is actually understand what is there in all these 100s of documents that come there. So first decision that needs to be made is whether it's really an appeal or not. Many times, you end up finding that what comes into the queue is not even an appeal. So that is determined by that agent, which is actually checking what it is. It actually understands the intent of the document.

It checks whether the document is complete. It checks if there are any missing records and missing documents that come in. Simple cases, let's say the document is not signed, and in a normal case, the appeal needs to be rejected. What this agent will do is go to the letter generation agent, and the letter generation agent will shoot out the letter saying, this document is incomplete, or there are missing documents. Please submit the right one.

So the two agents working together, completing the process. The most simple example I could think of in the healthcare scenario. We'll talk a little bit more about how the clinical agent works, how the payer policy agent works, and all that. But before that, we heard this mention of this word expedited appeals. Expedited are basically urgent appeals, which needs to be processed.

Technically, the time allowed is 72 hours, but most cases, it gets done in two to three hours. This could be because the patient's condition is deteriorating, patient is in pain, or anything else that could impact the patient or their experience of the member. So these are documents that come in. Again, like you saw, it could be one mention in a 100 page document. Now, how do we figure out that it has to be done; whether it's really an expedited or not? So if you can just continue the video, I'll just maybe show.

Page | 55

So what you see here is, you know, the agent is extracting documents as, that's what you see on the left panel. So it's extracting information from the image to know who the member is, the numbers. What is also critical is that a lot of this extraction is not just from an image. That needs to be validated with what's there in the legacy systems.

Ramesh spoke about it, how it's complex in healthcare. So you've got data sitting in multiple systems now. A simple example is, is a member eligible? That's sitting in some system that needs to be compared with the date at which the service is done, whether the person is eligible or not. Looks like a simple problem, but requires extraction from multiple systems there.

In the document, there is a big explanation. And what this agent has done is actually extracted why it thinks it's an expediter. So it's kind of given the verbiage which is sitting in the document. A nurse or the decision maker can just quickly look at it, see it gets highlighted in the document. In this case, we've taken a 7 page example.

But like I said, if it's a 100 page document, something gets highlighted. The nurse has to or the decision maker just has to look at this part. There is also a recommendation that comes. The recommendation where the AI is summarising the case, the agent is summarising it, you'll see it in the bottom of the screen. But really, it says why it's an expediter.

So this is actually a decision that's done. But because of compliance, it can't be given necessarily completely to an AI to do that, which is why there's still an expert that's there. And there is a whole lot of rules that get built into. On the right side, there are also some examples why something should be called or expedited. It just doesn't go with the verbiage.

There are a lot of policies to be considered. All of that, all those logics are also built into this whole system of Agentic that's handling this expedited case. Let's move on. Now, let's talk a little bit about a clinical example. And that's where I'll team with Krithika.

So we'll just take one example of somebody who requires a knee replacement. In this case, it's an appeal. They have not given that right document, but listen to this and I will just talk about it.

Video Presentation Starts

So what this agent does, I mean, in a simple case, if an ortho says, you require a surgery, we just can't go with that. So there are some guidelines that are mentioned there. You need to check whether the person is able

Page | 56

to walk, climb. Does the X-ray also support these scenarios? So there's a whole lot of guidelines that you'll keep seeing on the left. And what you see is the agent is actually giving the rationale to why it thinks it is, doing this. This person has osteoarthritis mentioned somewhere in the document. This person is not able to walk, mentioned somewhere in the document or the X-ray findings.

So all that gets automatically pulled from all these sets of attachments or documents that come, but still given and just recommending an option, there could be multiple mentions because we're talking about 100s of pages of documents, which is where the clinician or a nurse needs to look at.

But what we are talking about is saving significant time for a nurse to sit and do this work rather than have a person do that. So this is one part. And similarly, before a decision should be done, where other kind of treatments checked, where they're on painkillers, there was physiotherapy tried. So there are a lot of guidelines like that that you will get to see on the left side.

And Krithika will actually talk about what has gone into building this in the back end.

Krithika Srivats:

Can you hear me? Yeah, so basically, this has two parts, right? What you see on the green, that's actually the algorithm that's running in the back. And in the prior authorisation example, said a lot of times these policies are pretty clunky, 20, 30 pages long, sometimes much longer. And these are not in the form of questions.

So if you're reviewing the document, in no way can you come up with handful of these four or five questions. So there's a lot of sort of contextualisation of that policy that goes behind the backend. And then it gets converted into a decision logic essentially. Even there, that process in itself is sometimes very laborious. It can take about 100, 200 hours depending on how complex those guidelines are.

And we've used Generative AI to actually streamline some of that decision logic building to start with. But once you build that, the second thing that's happening is the Generative AI looking and making interpretation of these guidelines from the medical record.

So for example, if you see the first one here, it says, and such which conservative treatment have been attempted? First of all, it needs to understand what conservative treatment is. So we in some cases have simplified that at the question that says, okay, conservative treatment

Page | 57

could be physical therapy, some pain management, so on and so forth. So that is the decision logic, but then it still needs to understand pain medication.

In this case, it says Naproxen is prescribed here. It needs to understand that Naproxen is a pain medication. While that's not mentioned anywhere, it's just a treatment plan Naproxen. But the generative AI has then gone in using the SLM and the CLM that Madan spoke about, understood that Naproxen is a pain medication that is relevant in this case and so it then provides that as a recommendation. So these two things are working connectedly.

And then the third thing, although both in the case of appeal, UM or any other use cases where we've deployed this, the human in the loop component is the fact that this is simplifying that information. The recommended finger that you see is enabling the nurse to see where that information is. It actually sometimes gives a confidence score saying it's available in two, three different pages, but I have most confidence in this particular section of the document, whether that's in history and physical, laboratory tests, so on and so forth. And then the nurse uses that information to quickly validate, okay, did the lab test actually show this or did the x-ray show this? And then they annotate it and what would have otherwise taken easily in an appeal like this, it would take 30, 40 minutes. This can you shave off about 12, 15 minutes just from simplifying this process.

Ram Mohan Natarajan:

Thanks, Krithika. Let's move on. We'll just see one more agent in, in action. And what you will really see is not each of these agents acting independently. So this example that Krithika now spoke about will require an intake agent, may require a letter generation agent, will definitely require a clinical agent, may require a coding agent. So that's what you see on the screen. And some of these again in the back end, because it's pulling data from multiple systems will have its own agents which are able to fetch data from multiple systems there. So that's how this whole orchestration gets done.

And I will quickly show you a view of how it looks on the command centre. So like you see here, I mean, there are also limitations. There are a whole lot of stuff that needs to be checked. The entire rationale is given to help a person take a decision whether they should agree with this or they can overturn with that. So that's exactly…

Page | 58

Video Presentation Starts

Okay, we'll now talk about the one called the payer policy agents. So here again, the contracts between a payer and a provider are voluminous documents. There are discounts agreed for different line items, too. There are also payer specific policies that also need to be included when you're doing a reimbursement or when you're doing claim processing.

So there are, in a way, a lot of standard operating procedure documents that need to be referred to pay out a claim correctly or if somebody appeals, need to recheck that. So the whole process and why it gets expensive is also because you are reconstructing the whole claim again with this new set of documents. And then based on that, need to reprocess everything. Let's play. I'll keep talking while it's playing.

So here, again, the extraction document is taking pulling data there. Here, again, there is a lot of verification that needs to be done now with the core claim system. So you'll see some of those coming up too. There is a SOP document where that's there. You can also see thumbs up, thumbs down there. So we use a lot of this to train our AI to see whether it's doing it correctly. So that's where a lot of our algorithms and our clinical LMs are going in to get this train. And that way, the accuracy keeps increasing as we keep progressing. But that's not the end of it all, because models by definition can hallucinate. It can also drift. It can create some errors.

So how do we actually have that whole control system in place is also a critical element of how do you run or manage agents. I'll show you some screenshots of how we do that. Madan also had on the slide about knowledge retrieval there. So you can ask a question, about a document, or you can talk in some cases, and you get the answer there. So this, again, using a knowledge retrieval agent, which is actually pulling information to help a human in case they have to make a decision or they need clarity on eligibility or any other information from the documents.

Like I said, again, integrates into multiple back end systems because these are answers coming in from multiple systems for somebody to take a decision whether they should overturn or uphold the decision.

So here is a quick summary. So we did see multiple screenshots. But as AI matures in some of our processes, we start with this. We give a full summary. You don't need to see all the documents, 100 page documents that we are talking about, right? It's now extracted everything.

Hey, member ID exists in the claims platform. That's all a person needs to have checked. So the answer is here. And the rationale is it says, here is

Page | 59

the member ID number. Waiver of liability document is there. Here is where it is. Are they eligible for a claim? Is the date of service fine? All that is there in the screen. So this helps improve our productivity significantly.

But if they still have a doubt, they can go back to the document. They can go back to the screen. So we got those that I don't know if you managed to see. There were some icons on the top of the screen. So you just click that, get the whole summary, and then take a call whether you need to go back to the original document.

Yeah, let's continue playing here. So this is what I was talking about. So this is, again, another view of the command centre here.

Video Presentation Starts

So what you see here are screens. So you can find out the kind of exceptions that are coming in. This also talks about what we call as predictive ops you will see here, because the best part of all this is you're working with multiple data of thousands or hundreds of thousands of appeals that are getting done by all this agentic system. So one used to learn and train language models itself for this particular use case. But more importantly, for us to work on some of the stuff that Madan spoke about. He spoke about building predictive models.

Can we predict whether something is going to come back as an appeal? In which case, can we avoid the cost of a document coming back later by doing some things upfront? So it actually opens up a huge possibility in the world of analytics by making use of all the data that flows through these agentic systems.

So just to summarise quickly, we saw what Ramesh was speaking about as four different things that are different in a healthcare world. So one is you saw clinical documents. There are legacy systems involved. There's a huge amount of compliance and there are state regulations and federal regulations on what cannot be done or what decision can be taken by an AI that needs to be considered. So all of that are examples that you would have seen in this video. Thank you.

Over to you, Chris.

Chris Shiffert:

Good evening or good afternoon. My name is Chris Shiffert. I'm the Chief Growth Officer at Sagility. I'm your last speaker today before we have our panel Q&A. Before I get started, or go through my slides, a couple of things I want to point out. One, and I think you hopefully agree after hearing about all of our services and solutions today, mine is probably the easiest

Page | 60

job at Sagility, in that I'm tasked with connecting all these great services and solutions with our clients. And in all seriousness, it's a lot simpler than I ever thought it would be. That's one.

Two, you heard a lot about the pillars of which this company was built and how we go-to-market. One of them being our domain expertise, our industry expertise. And I want to talk about that for just a second because contextually it really helps understand the growth story. Healthcare is a big umbrella. I think you've heard throughout, the course today that, within that big umbrella, our slice is payer and provider and within that mostly payer.

And if you just look at the payer space in the U.S., there's maybe just under 300 payers out there. Maybe not all of them are candidates to be clients of ours, but most of them are. So it's a pretty finite group of prospects that we're going after. And we don't have to do lots of analytics around our prospective client base. Our services and solutions are diverse enough that we can service the biggest of the big, as well as the small and mid-sized markets equally as well. We've got a pretty diverse portfolio set and are pretty adaptable to matching those with the right type of health plan client. I think that was it.

Let me jump into my slides. Okay, so what we want to talk about today in terms of the growth story is really three things. One, at the beginning of our time today, Ramesh, pulled up some information about how we've grown this past year both with our large clients, our top five, as well as the rest of the business. The top five was about 10%, the rest of the business I think was closer to 30%. Great year last year consistent with how we've done in prior years. Why do we think we're going to continue growing at a healthy pace?

Then we'll get into the small markets, smaller and mid-sized plans I should say. A much higher growth rate, why we're bullish on the continuation of that trend, how that fits into our larger strategy. And then third, our new services, new offerings, some examples of what's in the works right now. Some of it you heard today. Those are coming both from internal R&D as well as M&A activity, but that's kind of the third leg of the growth story that we'll hear today.

And then before I started to get into those details, we didn't call it out specifically in the context of pipeline accelerators, but you did hear a lot about channel or partners that we're using, solution partnerships, technology partnerships, analysts and advisors, things of that nature. I'd say in the past we did a little bit of a restructuring a few months ago in a

Page | 61

way that has really unlocked a lot of activity within our company around working with partners, analysts, advisors. And we've gotten some really, really good traction just over the last few months. And I'll go as we continue through this portion of it into a little bit of detail about how that's manifesting itself with our pipeline and how that fits into our growth story as well, okay.

So let's start with the continued expansion of our top accounts. Ramesh had mentioned that there had been some conversation with this group or similar groups about, hey, we've had these big clients for years and years and years. They've continued to provide steady growth. How can that be, right? How can that continue on indefinitely? Couple that with a lot of the conversation that we've had today and a lot of the questions that we received today about the role of AI and how that is going to could cannibalise our revenue models and things of that nature.

So how are we going to do that? Here's what I want to talk about. One, we have this legacy that goes way back around 25 years of continuing to grow those relationships. That's way too long to ignore to think it's just going to end in one year because we have a new technology environment or the industry is going through some duress right now. So history and the historical context is certainly part of it.

More importantly and more detailed and factual is when you look at our and we'll take; we've been talking somewhat interchangeably about our top five clients as well as the seven of the top 10 nationals that we're working with. I think you're all familiar enough with the space to know who those big national plans are, and probably have some understanding about their buying behaviours and things like that. But what we have here is a depiction of the penetration of those top seven or the seven national plans that are clients of ours across our service offerings. And just to provide a verbal key, the blue dots indicate we have a presence in our practice area with these clients.

The solid, if it's really solid we have a stronger presence. If it's a little less hollow it's a little less so meaning there's room to grow there. The greens are where we're pretty solid in there. Maybe not a whole lot of growth potential in those particular practice areas, but most importantly the empty purple circles represent where we aren't doing anything with those clients in those practice areas and is really the growth opportunity that we have in the coming years.

So I think even our team is a little caught off guard that we really have one client where we are present in all five. Two clients where it's four or five

Page | 62

and plenty of growth opportunity in the rest. So that is one of the things that really excites us about continuing that growth story with our large clients.

And then lastly as regarding this segment of our customer base is we've had some real success with very unique different types of business models. You heard Roopam mention managed services and he gave a real live example of a different kind of a construct that's been a winner for us. We also have two contracts right now where we are helping our clients who happen to be national plans build out captive operations or build operate transfer types of arrangements where just by the nature of those types of deals, the relationship with those clients really changes. We're at a different, we're no longer in that vendor client relationship. We truly are partners. We're helping them build their future operations. And we see more room to grow with those types of deals.

And then lastly, we heard a lot about the MLR, the medical loss ratio types of services. Those as we move from the time and materials, basically the per transaction, the per hour types of billing arrangements into these types of constructs where we're delivering an outcome, we take on some risk, again, that takes you to a different level. We see a lot of promise and continuing to grow with those types of contractual arrangements.

My reflex is always to say do you have any questions, but I know I'm not supposed to do that, so I'll hold off on that. Okay. Two, the small and midmarket. So that table that we had in the previous slide, we kind of turned it upside down and wanted to show, how; the rest of our clients, the vast majority of our clients, we're really only in with one service line. I came from the BroadPath acquisition where we had roughly 30 clients. Of those 30 clients, maybe three had multiple lines of business when mapped to the Sagility portfolio. Tremendous growth opportunity with those. And that 30% where here we saw or 28.5%. We see this continuing certainly over the next couple of years as we get the story out.

And frankly, and I meant to mention this at the outset, as the industry is in this period right now where medical loss ratios are going up. The small and mid-market plans really, really, really need firms like us to help them even keep up with the nationals. So there's a lot of growth potential with those. I'm going way too fast I think.

And then finally strategic market expansion. We heard some of the story today about the Synchrony suite of solutions, really made tailored towards that mid-market, but it's going to be a growth driver for us. The market is ripe for it. We're counting on a lot for that. That is an example of

Page | 63

a build, coming out of our own R&D, leveraging a lot of partnerships, which also serve as their own accelerators. They've got their own growth organisations, their own pipelines. We support each other when it comes to that.

There's different flavours obviously of going to market with those, going to our existing clients with those workflow solutions, going to our partners clients with those workflow solutions, and then going to a net new for both of us. So it really does serve to accelerate the growth of the pipeline. Two that were acquired in terms of capabilities, the Medicare acquisition you heard earlier today that was one of the key offerings that came with the BroadPath acquisition. It was something Sagility didn't do. It's a new offering that we are now going to the Sagility book of business and expect a lot of success there.

And then before prior to the BroadPath acquisition, we acquired a company called DCI, which was an end-to-end payment integrity solution. Similar story in that kind of white space, green space for the Sagility book of business, also conveniently for the BroadPath book of business as well that we're going to continue to leverage those to get some growth. And then HEDIS and Star's is a last example of R&D driven innovation and new service lines, really focusing on the revenue side for our clients, making sure they're maximising their reimbursements on their government contracts and things of that nature. So I'm sorry. It was really fast.

And with 16 minutes left, I would certainly say, yes, let's take some questions, but I know my team is terrified, if I were to answer any. So maybe I suggest instead we start the Q&A, the general Q&A a little bit earlier. Is that appropriate? Okay. Is that good?

All right. So we're going to pull some chairs up. Do you want to hand one to me?

Question-and-Answer Session

Dhaivat Mehta:

Okay. So we had a few questions come in from the QR code, which was scanned. I would say most of them when we went through it were also asked by many of you when the mic came to you earlier. So we won't repeat those questions so that we can actually move on to more live questions from all of us here. Some of our leaders are on stage and you know others are here as well.

Page | 64

I'll request the mic volunteers to be bold and not be shy. So if the mic volunteers can be in the middle of the hall, Shiva, if the mic volunteers can be a bit more bold and so that we can make sure the mic reaches the person who wants to ask the question. I will step off the stage and if you can please raise your hand the mic will come to you quickly and we'll try to answer that question.

Roopam Narayan:

I think there were two questions which I had noted down which I didn't get the time to respond. One of the question was from you ma'am in terms of asking if the managed services case study was for the top five client or it was from there. So you probably saw it in Chris' slide, it's actually part of the top seven. It's one of those smaller LOBs of that client, and that's why it's a 25 million opportunity. So the way to think of it is that in a very large plan those LOBs exist where they cannot focus all the time on that and therefore they become the right candidates for managed services kind of outsourcing. So hope that answers your question.

I think there was a question from behind somebody had asked in terms of why people for managed services would choose Sagility versus competition. I think that was the question. Yes. So there is no right for us to go and demand the business as such. Competition is going to be there, competition is tough. They're equally good competitors in the market. The way to look at is in a managed services place where it comes is that if you're already there running that shop or running that business for the client, client has seen you perform. That's the first thing and that's where our delivery leader is here and we'll interact later on.

A very strong delivery track record, then Sri took about showed all those benefits we have consistently delivered over a period of time, 11.4 million of administrative saving, 5 million saving on late payment interest etc. All that we have already demonstrated. Now it becomes an easier discussion that if you give me this business where you've seen me deliver the day-to-day operation as well as transformation and now I'm going to risk my dollars, my money on delivering the result and transformation. I'm actually the risk mitigated partner for you to deliver those services. So that's the reason why we compete well in that kind of a space.

And it's always going to be, is that customer will create their own internal ROI or case of how much investment they need, what is the benefit they'll get internally at enterprise level and they'll compare with ours if we are giving better results, the business will likely come to us. So that's the case. Of course and internally also it's that their control points etc which we have to take care of in the customer side.

Page | 65

Somebody may want to give their business, somebody may not want to give their business, keep it in-house. So those kind of things happen. But roughly speaking, the reason why I made very clear managed services, we are specifically targeting where we have incumbency, where we have a track record of delivering is because exactly the point that you're bringing out. That's a competitive strength there, okay. So these were the two questions. I don't think I had any other question, which I have not answered in that time. So, we'll go from the new ones.

Siddharth Vora:

Ramesh Gopalan:

Yeah. Siddharth Vora from HSBC Mutual Fund. I'd like to ask a few initiatives which you will be doing for more cross sell of services. We saw a lot of scope in terms of top clients or mid and small clients as well that many of them are dominant on the claims management side, but other segments may be not working. So what is happening or it's purely a client maturity curve in terms of what they are accepting upfront and what they will do later on? More ideas or more initiatives in that?

Yeah, I'll start it off and then maybe Chris and even Roopam can add to that, right? So you're right, right? So if you look at the -- we've answered this question before. If you look at the journey of outsourcing, people took some of the more administrative aspects as the easy candidates for outsourcing and offshoring. So claims was one of those candidates that a lot of the payers chose as their first option to outsource. One of the things that we've spoken to you and hopefully came out today is clinical is a big growth area for us.

And clinical is in the relative scheme of things, it's a newer area, right? So most of the payers, I would say started looking at clinical not more than five to 10 years ago, right? Even our oldest clinical client is about probably 10, 12 years old. That's because there was the belief that the work was complex and the work couldn't be delivered by a partner and especially in an offshore kind of a situation, right. So and that's changed over the last 10 years.

We've proven to clients that we have the capability to find the right skill set even in some of our offshore locations. And we can deliver it to the same level of quality and outcomes that they can in-house. And the MLR being a big issue, like you heard today, people are now more open to even newer programs. So Krithika spoke about care management and so on. As long as you can show an ROI, people are willing to spend.

So it's not a question of I'm currently spending X and I want to reduce that. It's even if you want me to spend incremental, if you can show an ROI in a new program, I'm willing to spend that. So that's a completely new avenue

Page | 66

for growth. And then some of the capabilities were acquired capabilities. We were doing payment integrity, but we didn't have the end-to-end payment integrity capability.

That acquisition got us those capabilities. And payment integrity is an opportunity where clients are not constrained by how many vendors they have. So it's more if you can identify additional dollars, then I'm happy to give it to you because you may not be the first pass vendor, but you may be a second or a third pass. You can still identify opportunities, you get paid. So those new capabilities are so obviously things that we can take to our existing clients.

So that's why you see some of the white spaces, either because we didn't have those capabilities up until recently or even in the evolution of outsourcing, some of those areas are more recent than the others.

Chris Shiffert:

Yeah, I'd say I have one. The other if you looked at that chart and put, years of tenure by it, that would follow one, two, three, as well. In some cases, it just takes some time. Regarding the small and mid-market, there is still especially in that, we'll call it small and Roopam defined mid-market as a million and less, if small is 500,000 or 400,000 and less. There's still an awful lot of mission driven non-profit plans out there that really resisted outsourcing and still do to a large degree.

And they are now forced as I hate to use that term, but the per member per month admin costs of the large nationals and even that next level down is so low compared to theirs. They're almost at a point where they have no choice. They really need to do something to address those administrative costs just to stay just to keep up with the rest of the industry. So that's taken some time, but it's changing, it's actually changing pretty rapidly.

The MLR charts that Ramesh showed, those so yeah, we took the publicly available information and showed you. That dynamic exists throughout the industry. So just the confluence of those two things is going to drive a lot more people towards specialists like us to help take some of the cost out. Hope that answered the question.

Anil Nahata:

Thanks for a great set of presentations and demos. This is Anil Nahata of Parami Financial over here. I have a few questions. One is in terms of GTM, where you go to a customer who is a new customer to you essentially or one of the areas where outsourcing has not been explored much. What is the typical kind of a sales cycle and the time that you need

Page | 67

to convince a customer? And typically, how much kind of effort and do you need actual data to go along with the customer in such cases?

Chris Shiffert:

What was the last part of the question?

Anil Nahata: Do you need to actually demonstrate some data or some outcomes to the customer as a pilot or something?

Chris Shiffert:

Sure. So in those, there's a few a couple of ways to answer that. One, and this came through some of the acquisitions that Sagility has made with the Payment Integrity and BroadPath. Our solutions in the Payment Integrity example, they pay for themselves. It's an easy ROI and it helps that we don't get paid unless we find some money to recover.

On the BroadPath side, that company grew up solving very immediate niche problems for plans, very low barriers to entry, oftentimes very, very short contracts. And part of the acquisition was, okay, let's leverage that to get into more new logos. And that's certainly been a proven path of success. Then the last thing I would say is, I mentioned the accelerators, the advisors and analysts. Those personally are my favourite. If a plan, if a potential client has engaged with an independent, a third-party advisor, either be it one of the big four, somebody who specialises in healthcare. We're making sure that those firms know who we are and what we do.

They are in a very strong position to help a client you know define their problem, evaluate potential solutions, basically get about a third of the way through the sales cycle and then call us and certainly some competitors too. But all that really, really hard work upfront that takes a long time in a sales cycle and can be very difficult for a vendor to manoeuvre. If a third-party is doing that and we're jumping in the middle of the sales cycle, it compresses the time it takes to win that business and it's much more clearly defined for us.

Anil Nahata:

So you would say in the third case that you were explaining about with the third-party involved and all, it will be a factor of somewhere between six to 12 months kind of a thing or lesser than that?

Chris Shiffert:

I'm sorry, I didn't catch that. A sales cycle? If a third-party is involved, it's much shorter. Yes, closer to the six, right.

Anil Nahata:

Thank you for that. The second question is when the customers are looking at non which are not basically outcome based contracts, so you don't have an outcome based contracts and now you're having the agents coming. So there's a new way of looking at things from different from an

Page | 68

FTE kind of a thing. So how are the customers looking at pricing and how are you looking at pricing in those scenarios?

Roopam Narayan:

I think I can take that. So one is, which is usually the rebids, which are not very good for vendors, if you're not the incumbent rebids, your chances of success are very, very low. And we are obviously not in the business of cutting the FTE hourly rates and then going with the lower option. That's not usually not our area. So even clients who are going with saying that there's no managed services and we don't even have adequate data to give a managed services proposal, what we are providing is we're giving a roadmap, I mean in many places we have gone through a standard T&M based approach.

I mean, one client that we grew last year quite a bit, server managed this one when the process stabilised and we showed that they had 18, 19 claims per hour, we started doing 25, 26 claims. Then we said that look, how are we doing this? Why don't we convert into transaction rates so that we share the benefits together? So that's what was this one. And now we're negotiating and discussing other models which can incentivise both the organisations.

So yes, when there's a new client acquisition, we have to go by the process, whatever is the client's process, but we give them in advance a roadmap that look, we may enter like this and in clinical cases, even with large, very large payers we have done this is that we understand your problem, but we also understand your process that you're going to give us partial volumes, income competition with others, our rate card has to be in line with that, we are perfectly okay with it.

But that's not why we are focusing on this business is that we agree to take you through the journey into partial outcomes and then fully outcome based kind of model, which fortunately for us some of our competitors and many of the incumbents are not even trying to do to their own business. So we are coming as disruptors there. So the differentiator is now or later in future, you will still require a Gen AI infused outcome based model so that you can get benefited by the new technology which is coming in without you yourself having to take the risk and cost of implementation of a new technology.

Anil Nahata:

Thank you for that. And my final question is on the two segments that you are having, the broad segments, the payer and the providers. Of course, 90% of our business is the payer business, but you have a 10% on the provider side also. So how do you see this provider side going forward? And what is your focus on that business?

Page | 69

Ramesh Gopalan:

Yes, I'll take that question, right? So yes, in the interest of time, we didn't get into the provider business today. But I've taken up this question in the past. Providers is a segment that in the past historically, our focus is a lot more on the payer side, but we are doubling down on the provider side also.

If you look at one of our Agentic AI solution was focused on the revenue cycle. So today, we are looking at the process and that is more amenable to an Agentic solution because that's in a lot of cases, that's also currently an outcome based model. Providers, especially in the back end of the revenue cycle, they give us work and they pay us on outcomes, the cash that we collect for them. So we get the data from the client and everything else is dictated by us, starting with predictive modelling on what claims are likely to get denied, what claims we should follow-up at what point in time and so on. And so all of those point solutions are already there.

Now we are doubling down and saying, can we create an Agentic layer on top of that and bring a lot more efficiencies to that process. So yes, that's also in the same roadmap with respect to technology and transformation. And we intend to double down on the provider business. And some of the other opportunities that we're also looking at providers, some of the solutions that Krithika was talking about in the clinical space, right? Today, a lot of what we do in provider is the traditional RCM based services. But we are also trying to take some of the capabilities that we've built on the payer side with respect to some of our clinical solutions and trying to take them to the provider.

So from both those aspects, we are likely to focus a lot more on the provider and hopefully accelerate the growth on that side of the business too.

Atul Mehra:

Chris Shiffert:

This is Atul Mehra from Motilal Oswal. Thank you once again. I have a question for Chris. If you had to grow at 2x the growth that you are currently growing at, what are the key constraints in terms of either teams or time or if you can explain with some examples? Thanks.

It's less of a constraint than more of a what's the business mix, right? Some of our offerings implement much faster. Some of them take a long time. So growing at a rate like that is certainly doable for us. We have a very scalable model, our operations are, just the global footprint that we have that removes a lot of constraints that certainly I was used to at BroadPath.

Page | 70

But I'll go back to that my initial answer. It really depends on the business mix as some are harder to implement, take longer to implement than others.

Atul Mehra:

Chris Shiffert:

But in terms of like you showcased in your presentation there are so many low hanging fruits in terms of cross sell, in terms of the new logos you'll have added etc. So typically when it comes to a client conversion, to like from a conversation to a conversion. I think there was question previously as well in terms of the lead cycle it takes and so on. So what is like in the conversation to conversion, what are the key roadblocks you all face when a conversion doesn't happen?

So because a lot of times it's the first conversation they've had on some of these topics especially with those smaller plans, right. Nobody's approached them before with a, hey, we've got this clinical service, here's the value it provides. So you are initiating the sales process as opposed to, if they've identified a problem internally, organise an initiative to go find a solution for it. So that's where we are with a lot of the, especially, the smaller clients where there is that green spaces.

We're introducing that, maybe they thought of it before, but they never did anything about it. So it just takes longer that way when we're the ones that are the -- we're the impetus for picking up the initiative, not them.

Atul Mehra:

• Chris Shiffert:

Ramesh Gopalan:

And if there was one thing you could do, which you are not doing currently to further accelerate growth, what would that be?

One thing, you want one thing?

Yeah. I can talk of one other things, right? One of the things that historically we didn't do well and like Chris said, we've started focusing a lot more is the partnership ecosystem. So thanks to both Roopam and Chris, and another teammate of ours. We've started focusing a lot more on the partner ecosystem. It does two things. One, it enlarges the scope of what you can deliver. The Synchrony suite of solutions is something that we couldn't be delivering on our own. We need partners for that.

And two the point that Chris made, the minute we start working with a partner, it also kind of opens up other opportunities that you didn't have because they have their own pipeline. So that becomes part of your pipeline now. And so you can not only take their solutions to your existing clients, but it also opens up another channel of pipeline. I think that's one thing, I would think, we didn't do so well in the past, which hopefully will result in more deals in the future.

Page | 71

Baidik Sarkar:

Ramesh Gopalan:

Thank you. Ramesh, hi. This is Baidik Sarkar from Unifi Capital. I think what my industry colleague was trying to ask you was what kind of growth rates are you willing to commit to yourself and us today evening. So if you could please address that for all of us, point number one. And my end of the question is, look, we've been acquisitive for the right reasons. But like Chris said, there continues to be enormous amount of white spaces. So given that we're very close to deleveraging event in January of 2027, our cash accretion is what it is, how close are we to another M&A event? And what white spaces are we looking at? And financially, what kind of ROIs do you expect from that given sales cycles given the valuations are? So if you could just marry these three answers into a simplistic narrative.

First of all, Atul is very open. So he can ask, if he really wants to know the growth rate, he would asked it specifically. But yeah, look, specific guidances we'll give with our earnings call that will happen in May. But broadly, given all that you heard today, we believe that there is no reason for us not to grow at historical growth rates. So that's the broad guidance I'd give. While there's been a lot of questions around AI and its impact and its disruption and so on.

There's nothing internally that we believe will constrain us from continuing to grow at historical rates. The questions around can you accelerate it further and so on are good questions, and obviously, we would also like to grow faster than we are growing today. But at least one thing that I can say, what we see, at least for the near term, I don't want to make a five year projection, is we will continue to grow at historical rates in the, like I said, low double-digits to mid-teens kind of a number. But specific guidances for the next financial, we will give closer to the earnings call.

Your second question, yes, M&A has always been a strategy of ours. And like I've said, historically, two reasons we do M&A. One, the BroadPath kind of M&A was more to get access to clients. And like Chris said, we have about 30 odd clients where we just do one service today, opens up a huge opportunity to cross sell other services.

And the second thing is from a capabilities point of view. Unfortunately, there are not a lot of BroadPath like targets, especially focused on healthcare alone, because that, I mean, we'd always love to shorten the sales cycle. And if there's an established relationship that makes the job much easier. But there are not a lot of BroadPath like targets available.

So most of the acquisitions are going to be more capability focused. And as you saw in the presentation today, we are emphasising a lot more on

Page | 72

our clinical solutions, right. So there's going to be, I mean, there are a couple that we are evaluating, and I can't comment on how soon or how longer it will take to close those. But if we were to do an acquisition, the ones that we are looking at currently are more in the clinical domain.

And yeah, ROI from an acquisition is obviously, as part of the business case, we anything that we pay, the multiples we pay compared to our current multiples and what's the accretion or dilution, all of those get into the equation. Most of the times, I mean, from an ROI perspective, acquisitions have worked for us, at least the recent ones that we've done. And you can rest assured that we'll pay the right value and make it accretive in the long run.

Chetan Shah:

Hi, this is Chetan Shah from Jeet Capital. Just continue to what Baidik is trying to understand and something add to that. Can you just give us some idea about you like we have spent plenty of time on AI and AI related thing, whether it is a disruptor or an enabler or whatever it is all about. But can this automation of business implementation to our clients, thanks to AI and AI related activity, increase the size of the opportunity for us, one part that.

And second, some of the clients who are at the sideline as of now, either due to the cost, second due to any other reason can come into the fold of our opportunity space.

And third and most important which some portion Chris answer is that sales cycle is little longer. One of the reason is the implementation of some of the modules of our businesses takes time and this can AI enable that to do it at a much faster pace and suddenly the client will be more than happy to get into the opportunity without disrupting his activity or ongoing business? These are the three.

Ramesh Gopalan:

Roopam Narayan:

I'll try and remember the three, anybody wants to pick up. Roopam, you want pick up any of that?

Yeah. I can take that and same thing probably you'll have to help me with the parts of the question. AI is definitely speeding up implementation and time to revenue for us. But there are certain cycles that you have to be clear of, like Medicare life cycle product that we talked about. AEP does not start till October. So even if we go sign the deal now, it's not going to. And really speaking, the strength is going to come in next calendar year, because the new plan benefit filing or new benefit design that the client will do is actually for calendar year '28. So it's really we have to time our revenue projection based on when these new services will get consumed

Page | 73

by the customer. And that happens based on the cycle of the business that goes on there.

Similarly, things like payment integrity. Even when we implement everything and start recovering or start showing which are the claims to be recovered or something, till the time recoveries actually happen, we don't get the revenue. So there is a cycle for the implementation, the readiness of the client side to flag those identified claims, review that and say, okay, now you go recover and then recoveries happen. So there are certain business cycles that we cannot speed up even with Gen AI or without Gen AI because those are business cycles, not operational cycles.

But then the other part of where the implementation, the IT services components and all, definitely there are lot of accelerators have come into the market. The testing has become much simpler, faster. So those things are taken care of. But they are all, they are just allowing us as to be right on time, quality and costs and be delivered in time.

The real cycle is the business cycle, which is linked to AEP, linked to claims recoveries, linked to claims processing actually starting. Most of the cutovers that happen when they change vendors or something is data service based and that is because health plans have accumulators in your annual this one is that your total service that you have deductibles, etc., is by calendar year. So a lot of those plans will not start with a new vendor on -- with the date of service date of anything other than January 1. So these are some of the constraints that we cannot change, okay? So that's really on the cycle.

The other two questions, if you can just come back on.

Ramesh Gopalan:

Yeah. First question, I mean, hopefully, I'll answer it, but hopefully throughout the day today, that's what we were trying to address. The one message that I want all of you to take over is we are not viewing Gen AI as a disruptor. Yes, it is going to disrupt, but we are viewing Gen AI as something that's going to expand the scope of what we can offer and accelerate our growth. So if that's one message that I want all of you to take back today is that message.

To the question on how are you protecting your revenues, we don't want to use the term of protecting our revenue, because we are in the business of adding value to our clients. We are not going to stop generating efficiency because we have to protect our revenue. Our job is to generate

Page | 74

efficiencies and pass on those efficiencies to the client and make the relationship stickier so that we can continue to grow with those clients.

So with that in mind, so we are not using the term protecting our revenue. We are using the term, we will embrace Gen AI, we will generate those efficiencies, we'll give them back to the client, and we'll use these new models to go to our existing clients and other clients and win more business, right? So that's -- and you've seen with a number of use cases how we can enlarge the scope of what we do. Managed services, Roopam gave the example, a current 6 million can actually become a 20 million, even though, I'm committing to a 5 million cost takeout.

There are models like that. We spoke about MLR, but it's not a question of cost of service reduction, but it's more a return on investment decision. So there are a number of examples we've given today to kind of dispel the notion that you have to view AI purely as a disruptor to your revenue. But hopefully the message we are trying to deliver is AI can also accelerate our growth.

Chetan Shah:

Ramesh Gopalan:

Yeah, thank you so much. And just one very specific question. In of the clients and this is completely my personal assessment speaking to some of your existing or potential customer. One reason why the things get delayed is internally within the top or the mid layer team, is not ready to bring such technology into the system because some of them may become redundant within the organisation. Is there also one of the reason why, this is what we are getting the feedback. I'm not sure about this, but I'm just want to clarify that.

I'll start and I'll ask Madan or someone else to right, which is very true. In healthcare in general, you can probably know this, that adopting newer technologies takes time. And hopefully, one of the things that we also told you is, one, they are slow in general. Two, also it's the nature of the industry. The regulatory complexity, all of the data protection, data privacy, all of those things generally are reasons why clients take more time to make sure that they don't do something that's going to cost them millions and billions of dollars to fix.

So a lot of times our use cases, while we have the capability to deploy some use cases, we are waiting for clients to give us permission to do that, because we are still working on the frameworks for the guardrails and how they -- the whole framework for deploying AI. Madan you wanted to…

Page | 75

Madan Moudgal:

Yeah. So one of the things that to make AI effective and impactful, you need to organise your data. And that can be time consuming if you have not already taken the time to, for instance, build to do the data aggregation, build your data lakes and such that allow for this curated data then to be made readily available, it can be less than the impactful.

So sometimes when we talk to our clients, they're busy fixing those things, and so they may want to delay or put us off for a period of time. There's still ways in which that we can bring about some change and make an impact, but we have to work around those constraints. Gen AI by itself can be an accelerator there, but there also as Ramesh pointed out, some amount of trepidation around, well, how is this going to expose us to certain risks? And then there are governance committees that get involved. They want to have their say.

So there's all of these types of gates that you have to cross before you get to the point of saying, okay, we get it, now we can move forward. A lot of times, we will get the buy in from our business partners on the client side. The IT partners sometimes will take longer to come around, and give us the go ahead. So it's just the nature of the process, and we need to work through it.

Namit Arora:

Thanks. Good evening, everyone. This is Namit Arora from IndGrowth Capital. Firstly, many thanks for putting the day together. It's been very useful to have the entire leadership team here. I have one question around competitive landscape. How do you look at that and there were a couple of interesting slides, one was where you had your ecosystem of partners, for example, you even mentioned CTS and others.

And the second, somebody mentioned that even when clients are looking at innovation, they've actually asked you to coordinate with several other players who may be your competitors. So give us some colour on how you think of competitive landscape for you, number one.

And number two is, as some of the larger other companies, let's say, like Accenture, IBM or Infy and others. I mean, I know your business is very, very specialised. You have domain knowledge. But as they sort of feel some pressure with AI and others, are they trying to make some inroads here? Just give us some thoughts on all of that. Thank you.

Ramesh Gopalan:

Yeah, I'll start off and Chris, Roopam can add to that, right? Look, in the space that we are in, which is, again, we define it as transforming healthcare operations. We come across a set of, I would say about five, six competitors in most scenarios, right. But the market is broad enough,

Page | 76

I mean, if you look at the service lines, the service areas, the domain, it's large enough that there's ample room, in my view, for everybody to grow.

More importantly, the things that differentiate us from any of them is what I started the session with. Our domain expertise and our broad service portfolio are two things that we think are superior to most of our competition. And being focused on one industry has its benefits. Obviously, your market is restricted to that one industry. But everything that we do today is focused on solving business problems in that industry.

So a lot of our investments, be it in AI, be it in other tech solutions, is all extremely focused on solving business problems that most of our client’s encounter, right. So that gives us the razor sharp focus as well. And that's always helped us deal with competition. But to your point, yes, it's not a huge list of competitors, but we typically see the same five, six names when we compete.

Roopam Narayan:

Yeah. And I think you named names, so I'm not going to repeat them, but I'm just saying is that those vendors are already competitors for us. So it's no different from what's happening today. Yes, if they are pressured, they like to expand in the space we are in. But it's the same thing. They'll have to fight against our strong delivery, our incumbency, our relationships, our transformation that we are already bringing in. So we don't see that as a big threat that something will happen, because they have pressure on the IT services side and they'll come in, because they are already there and we are competing well against them.

And we have respect for those organisations, they'll definitely come with their strengths. But we have our own strengths in places they'll win, in places where we are winning we'll continue to win. So I think that ecosystem and that competitive dynamics is not likely to change.

Ramesh Gopalan:

Namit Arora:

And the other thing I want to point out to in addition to what Roopam said is, our pricing constructs that we spoke about. So we are not going to market as a software vendor. We are not charging license revenue. Yes, I mean, I'm not saying we'll charge zero implementation, but we can absorb some of those costs and amortise it over a period of time. So we are not looking at generating upfront revenue from some of the transformation that we do. So that's also a big advantage that we have compared to others who want to monetise it right up front.

Thank you very much, Ramesh, Roopam, and the entire team, and all the best to the entire team. Thank you.

Page | 77

Pallavi Deshpande: Hello this is Pallavi Deshpande from Sameeksha Capital. The SaaS tools, I mean how has that uptick picked up? And in terms of our ability, our own, how many engineers would we have writing code?

Madan Moudgal:

Is your question more about how are we using Gen AI to improve the productivity of our own engineering stuff?

Pallavi Deshpande: Madan Moudgal:

Right.

Okay. Yeah. So certainly, we have jumped in with both feet. We clearly understand the power that it gives us across the entire not just in from the standpoint of software coding, the entire software development life cycle. So there is an internal initiative that's underway. It started about a year ago, and we're coming up to a point where we're having our first measurement of the impact on that productivity gains. And the data so far is quite promising. So we will continue to double down on that.

We certainly see the benefits. At the same time, we are also, I would say cautious because it's a double edged sword. If you don't properly monitor and supervise the use of these tools, it can lead to some, not just bad practices, but it can also lead to errors entering into your code base. You have to be very careful about that because what we obviously don't want to do is to cause any of our clients to lose confidence in our abilities. So we are very carefully, but enthusiastically taking advantage of these capabilities.

Ramesh Gopalan:

And the other reason is we don't build these out to clients directly. So for us, any efficiency that we can generate is very important. So we have no hesitation in using it to the fullest to generate whatever efficiencies we can.

Roopam Narayan:

Just to complement Madan's team with the similar workforce type, the value creation projects that we are doing for top three clients, now we are doing for six clients. So obviously all the additive knowledge that we have built that Madan was showing those cumulative knowledge that also plays and also the productivity gains that we have with all these tools which are coming in.

So from with the similar cost structure, we are able to handle much larger client base for value creation projects.

Pallavi Deshpande:

My second question would be what would be like you showed us different pieces of examples here or there. So like if I look at it right from the typical, right from the start to the end, member enrolment and entire billing.

Page | 78

Which pieces, or how should I say how do you imagine it with AI two years hence in this whole flow, where does which pieces get impacted most?

Ramesh Gopalan:

I'll give a broad answer and Roopam will probably add to that. We think every single process that you saw in our process thing can benefit from the use of AI. So I wouldn't say there's anything out there where AI can't add value. Like Madan said, even within and those are all broader functions and processes. Within that process, the tasks at a task level, we may not deploy Gen AI across the board.

There are some tasks that can benefit through other forms of agents. But at a process level and at a function level, I think every function can take the advantage of AI. But in terms of the impact from an efficiency point of view, obviously processes which consume a lot of unstructured data, we'll see a lot more benefits coming out of AI. Krithika and Ram showed the example of documents, 10 page, 50 page text.

Those kinds of documents when somebody has to manually go through them and make decisions and so on, versus the use of AI to at least put together the recommendation before the human reviews and makes the final call. The level of efficiencies that you can generate are far higher. Then a process which is already rules based on very structure. So that's the two ways I would answer the question. If Roopam, you want to add anything?

Roopam Narayan:

No, I'll just bring back the same point. We are not coming with a mindset of trying to see what's AI proof or reduced AI disruption, because that's a losing game in the long run anyway. So think of it as AI is going to penetrate every part of the operations. Maybe some places your regulatory issues like anything which is about denial of access to care or delay of access to care or an adverse decision that you're making against a member or a provider. These will certainly require a lot of human involvement for sure.

But other than that, you will have AI. Now there's always going to be a degree of effectiveness of AI, a process which is using a lot of structured data, a process which has limited volumes in each type of, let's say, claims spends. There are hundreds of types of claims spends with very low volume. Probably it will not be the right case for just going after Gen AI or reading SOPs and trying to solve this immediately. So we'll have those kinds of things which are coming up.

What we are looking at is assume that Gen AI is available as one more means of delivering the business process outcome. Now come with a solution which can handle and deliver client the outcomes. So our

Page | 79

thinking is different from the productive mindset of saying is that let's go focus in areas which will not be impacted by Gen AI because that's not in my opinion, that's not a sustainable way of growing the business. I mean anybody else wants to add to that? Ramesh Gopalan: I think, yes, we've made that point. Ravi Menon: Hi, this is Ravi Menon from Macquarie Capital. Could you give a breakup of your payer revenue by service? That's how much would be claims, how much would be from Synchrony, how much would be provider that credentials validation stuff like that?

Abhishek Kayan: So I mean our payer revenues are around 90% of our total revenues. Onethird of our payer revenues are claims revenues, one-third of our payer revenues are your clinical and payment integrity and the other one-third is a member and provider life cycle. Ravi Menon: And how about the payment integrity and all that? Abhishek Kayan: Payment integrity and clinical combined is closer to 30%, one-third. Ravi Menon: And Synchrony itself? Chris Shiffert: Synchrony cuts across, so we don't track it like that. We track things according to those service practice areas and Synchrony purposely cuts across all of that.

Ramesh Gopalan: And from a capability evolution like I showed. So the first three pillars are what got us today, and then the automation of complex workflows and Synchrony, those are more forward looking. So those are solutions that we are taking to market currently for future proofing our business.

Ravi Menon: And although you're very healthcare focused, I mean, how about adjacencies like say, in retail, I've seen that some of your competitors have, something for similar to the sort of leakages that are there and payments, you have solutions that could be addressed there?

Ramesh Gopalan: No, at this point, we want to continue to remain a healthcare firm. Like I've mentioned in the past, we are thinking about relooking at the definition of healthcare that we've focused on in the past, which has been restricted to payers and providers. We might expand upon that definition and look at adjacencies within the healthcare space. But that's the extent to which we will go, at least in the foreseeable future. So we don't intend going to retail or financial services and so on.

Page | 80

Varun Gandhi:

Hi, I'm back here. Varun Gandhi from Finavenue Growth Fund, back here. So my question is and it's trying to understand the core of your business model. So we work with healthcare systems like Availity or Health Access. But Sagility's position, so we act as a traditional BPO provider by bringing in our clinically trained workforce and trying to orchestrate workflows. But beyond that, what's the sticky element that makes the customer stick to you? Is it the execution prowess that you demonstrate?

Just trying to understand what's the proprietary element here that Sagility brings to the equation?

Ramesh Gopalan:

I'm a little disappointed that after five hours, you're asking me that question. Anyway, so I mean, don't mind repeating. If we don't, if you look at me and my colleagues in the five hours, we didn't use the word BPO. Yes, broadly, you want to classify us as a BPO, yes, we are a BPO. But we are not in the business of clients defining the rules and the process documentation and we following the documentation to deliver an outcome.

We built this business over 25 years, and today we believe we have the domain expertise to redesign processes to give clients business outcomes through changes that they haven't been able to implement themselves. And if, yes, if that's something, is something that we've done, like I said, over 25 years, and that those capabilities is what differentiates us. To the question on competition, yes, there are other people who operate in the healthcare space. We differentiate because of the domain expertise that we bring in.

And all of our transformation capabilities are focused on solving specific business problems that our clients have. And that's what helps us win business and grow the business. So I don't know if Roopam or Chris want to add to that. But today, a large part of why we win is the fact that we have the deepest understanding of how the ecosystem works. We also have a very good understanding of our clients' ecosystem and what kinds of solutions are most relevant for our clients.

And so it's not always, order taker mentality. So it's not that the clients determine, what we should do at what point in time. Yes, the relationship might start that way, they have a need and they call us in, but the evolution of that relationship is we proactively taking solutions to them and solving their business problems.

Varun Gandhi:

Understood. My second question would be, you showcased many AI products that we have, but these will sit as wrappers on top of these

Page | 81

healthcare systems that I earlier referred to. And so my understanding is that these products would largely be platform agnostic. If so, how susceptible are we to the risk that the healthcare system providers or these guys who are owning the tech stack, they tomorrow develop their own AI products to or Agentic AI products who are orchestrating the same workflows, then there's a good chance they may displace us. So some comments over there.

Madan Moudgal:

Yeah. I'll start by saying that what we take to and make available to our clients is not just the technology. It's the technology and the services together in a lot of cases. So the entities that you refer to as the health system I think you call them health system. I think you mentioned health equity and people like that. They're offering a technology by itself. So, yes, you're right that over time, you'll -- they'll embed more Gen AI. We expect them to. And we expect to take advantage of the Gen AI capabilities that they've embedded in their platforms.

But there'll always be room for us to add additional value, because we combine those features with our additional services in order to deliver a full-fledged offering to them. That's of more value to them than just the technology by itself.

Varun Gandhi:

Madan Moudgal:

Ramesh Gopalan:

But then a lot of the value capture would happen on the tech layer, rather than the services layer? Just trying to understand.

I would say the value capture happens in the combination of the two. That's why we're delivering the value to them, in my view.

See, we are not going to market saying we have the Gen AI capabilities. The contracts that we are signing up with clients, we are still accountable for the business outcomes. That's the biggest difference. We are not going to, and Gen AI or any other solution is an enabler to delivering the business outcome because my clients are still holding me accountable for the business outcomes, irrespective of whether I use 50% automation or 80% automation.

So in the whole equation, whether it's and a lot of the platforms and Synchrony that we are talking about, the orchestration layer could be done by us, could be done by our partners, could be done by a combination of us and our partners. But ultimately, the client is seeing value in what's the cost of delivering that and whether you're still committing to the same business outcomes.

Varun Gandhi:

Got it. So you're trying to say that the value is in the outcome rather than the process?

Page | 82

Ramesh Gopalan:

Varun Gandhi:

Roopam Narayan:

Yes.

Thank you for that.

I think the other way of looking at it is, I think you mentioned Availity, Health Access, a few of these kind of companies. So everybody has a Gen AI play, but it's restricted to their platform. And the platform is not an endto-end process, okay. The Synchrony Medicare life cycle that we talked about, think of where it's from the design of the benefit products to filing to that benefit information moving into the sales system, the consumer portal, and somebody making sales calls, brokers handling the same information, then you're enrolling the member, the same benefit information has to move into your premium billing and member enrolment system and into the claim system.

These are five, six different systems and everybody has Gen AI play in their own. But if I have to deliver this, which we are delivering as an outcome from end-to-end and that includes even paper fax enrolments which are coming in and which import and output is print out where you're sending out ID cards and everything is going to an external player. There's none of these platform vendors can have a Gen AI play across the entire life cycle. That's where we come in with an outcome-based solution and say that we have our own Gen AI agents, for example, fax intelligent content processing is with us, output management is with us, orchestration layer is with us and then we are calling and we are working with the Benefit One products, GenAI module.

So it's something like that you're basically taking the entire outcome, then coming in how you can deliver optimally with the each one of those platforms and utilising what's the core automation that has been built into the platform.

Varun Gandhi:

Roopam Narayan:

Varun Gandhi:

Roopam Narayan: Ramesh Gopalan:

Effectively, you're saying that you're like the connection point between these different layers of system that we have?

Connection point, orchestration point, responsibility of human in the loop, outcome delivered to the end.

Got you. Thank you very much and thank you again for hosting us today.

Yes, thank you.

I think we are out of time. Any further questions, we can take your last question.

Page | 83

Participant:

I have a broad question on margin front and our employee headcount, like is there any possibility that as we are wrapping up the AI layer in our existing offering, client may ask for the benefit pass on in future?

And second, our margins are more than 22% or 25% year about. Is there any chances that it may face some sort of compression going forward? And on headcount front, you consider that going forward due to more and more AI led offering, we may reduce our headcount significantly considering the current strength of near about 40,000 employee workforce?

Ramesh Gopalan:

Yeah. It's difficult today to give specific numbers. So but broadly, directionally, what we can say is, look, as we deploy more and more of this AI and Agentic AI solutions, if you look at our revenue productivity at a geo level, that is likely to significantly improve. Our overall today, headcount and revenue per FTE is also determined by the geographic mix. We have people in the U.S. who generate higher revenue per headcount. We have people in India who generate much lower revenue per headcount, right? So that geographic mix is still going to play out in the future.

But if you look at like-to-like, if I look at the revenue per FTE per employee in India today versus what it's likely to be two years or three years down the road, it's likely to show a significant movement upwards. And then at the enterprise level, it will be a weighted average of what headcounts I have. So broadly, yes, We are confident that we'll be delivering more revenue per headcount at a geo level and depending on the business mix, it will also show up in the overall company level.

With that and with the growth rates that we are expecting, whether headcount will stay static, whether it will reduce or it will increase, but not increase at the same proportion as revenues, those are numbers we need to play out. But otherwise, at a geo level, like I said, you will see a significant improvement in revenue per headcount.

Participant:

Ramesh Gopalan:

On margins?

Margin, again, is a question of how we want to grow. So broadly, in the past, we've said that we are comfortable at the current margin levels we are in, while automation, transformation, Agentic AI could help us improve margins, we want to plow that back as benefits that we pass on to clients and continue to grow our business. And that broadly still remains the same. All of these tools will enable us to generate better

Page | 84

efficiencies. It's a question of what portion of those efficiencies we pass back to clients versus what portion that we retain for ourselves.

And broadly, we are happy with the margins where they are, so we'll try our best to retain these margins, while utilising the benefits that we generate to help us continue to grow. So that's been the broad strategy even in the past, and we'll continue to do that in the future as well.

Participant: Thank you.

Pallavi Deshpande:

If you look at outsourcing penetration, 20% or 40% in your top five clients. What is stopping the faster growth in the industry basically. I mean is it the client itself being hesitant to outsource this, and so then..

• Chris Shiffert:

I didn't understand the question.

Ramesh Gopalan:

Overall outsourcing penetration is low, what is stopping?

• Chris Shiffert:

Overall across the industry is low. I mean healthcare has traditionally been the most conservative industry in terms of embracing outsourcing. A lot of that is regulatory driven. But I mean it's also, as always said healthcare and financial services are the most personal industries out there, because they, I mean, you're talking about your health and your finances, the two most personal things to you. Companies are always, they position themselves as the trusted entity in that interaction. So it's just natural that healthcare is more conservative that way.

But another way to look at it is we had a meeting a few months ago with one of the large consulting firms and they brought up the question of what really is a health plan? And it was a conversation that this gentleman said that he had been having with his health plan clients where they're really redefining what is core to their business. I mean 10 years ago you would have said clinical operations was core to their business and now that's the fastest growing segment of ours.

So I think health plans are also redefining, they're having that conversation constantly, which I think will lead to more outsourcing as they define what's really core to them, because they keep siphoning off processes and farming them out.

Pallavi Deshpande:

And lastly, how open would clients be to rewrite their processes to take advantage of Agentic AI?

Ramesh Gopalan:

Roopam, do you want to answer that?

Page | 85

Roopam Narayan:

I think most of them are very open. But I think if somebody also brought in this thing of that contention on the legacy IT platforms or something. So think about it as a business organisation. You will always go for projects which will give you highest return and quickest returns. So many of them having legacy IT platforms is a big drive already happening where they're changing their legacy systems, because the IT cost itself was very high. So now you have, I mean while we're not saying that we are protected because of that, but that's also true.

I mean, if the saving of Generative AI and platform change is higher, is delivering higher returns and quicker returns, then the projects which are operational projects to implement Gen AI and reduce operational costs will take the backseat. So you will go with the higher ROI projects first and it's happened in a couple of cases for us where we have a clear path for the ROI and client has a very clear direction that they will use that in the Phase 2, Phase 1 for next two or three years as they are undergoing a platform change, because that's giving them $10 million, $12 million annual saving. And ours is giving them $1.8 million savings. So there's no competition when that happens.

So everybody is extremely keen to bring in Gen AI, but at what point of time depends on other competing projects which may deliver higher our ROI and quicker ROI.

Pallavi Deshpande:

Thank you.

Ramesh Gopalan:

So thank you again for everyone for coming. So we will break now for drinks and dinner, but we'll all be available outside so we can continue the conversation. Thank you.

Dhaivat Mehta:

Thank you, Ramesh. Thank you, everyone. Thank you for those amazing questions. I must thank our partners, CDR, Keystone and many of the teams that helped internally make this happen. All the teams at Sagility. Thank you once again. Let's continue this over dinner and drinks. Thank you.

Disclaimer: This is a transcription and may contain transcription errors. The transcript has been edited for clarity. The Company takes no responsibility of such errors, although an effort has been made to ensure high level of accuracy.

Page | 86