Guest Speaker:
Transcripts
Welcome to Tech Lyceum. This is a podcast from Birlasoft. My name is Neerja, and this is a place where we look at the technologies that are shaping enterprise transformation.
In today's episode, we ask a timely question for insurers: Will Agentic AI redefine the insurance core? Now, core platforms such as Guidewire, Duck Creek, and Majesco continue to power critical insurance operations, but as agentic AI evolves, the focus is shifting from co-modernization alone to how AI, humans, data, and workflows come together to create a smarter operating model. So there's a lot to unpack here, and to help us do just that, I'm joined by Mradul Kapoor.
Now, few leaders understand insurance core transformation as well as Mradul Kapur. With over two decades of experience, Mradul has helped tier one insurance carriers modernize complex core platforms across Guidewire and Majesco, managing technology portfolios of over 30 million dollars and turning transformation programs into measurable business impact.
At Birlasoft, he leads North America insurance technology transformation and delivery, helping insurers move beyond modernization to unlock AI-led workflows, operational efficiency, and smarter operating models. Mradul, it's great to have you here to discuss this with us. Thank you for lending us your expertise.
In today's episode, we ask a timely question for insurers: Will Agentic AI redefine the insurance core? Now, core platforms such as Guidewire, Duck Creek, and Majesco continue to power critical insurance operations, but as agentic AI evolves, the focus is shifting from co-modernization alone to how AI, humans, data, and workflows come together to create a smarter operating model. So there's a lot to unpack here, and to help us do just that, I'm joined by Mradul Kapoor.
Now, few leaders understand insurance core transformation as well as Mradul Kapur. With over two decades of experience, Mradul has helped tier one insurance carriers modernize complex core platforms across Guidewire and Majesco, managing technology portfolios of over 30 million dollars and turning transformation programs into measurable business impact.
At Birlasoft, he leads North America insurance technology transformation and delivery, helping insurers move beyond modernization to unlock AI-led workflows, operational efficiency, and smarter operating models. Mradul, it's great to have you here to discuss this with us. Thank you for lending us your expertise.
Listen On
Speaker – Mradul - 01:44
Thanks a lot, Neerja, and thanks for bringing that energy again. Appreciate you having me here. So, yep, all set for the discussion.
Thanks a lot, Neerja, and thanks for bringing that energy again. Appreciate you having me here. So, yep, all set for the discussion.
Speaker – Neerja - 01:51
Super. Let's get started. And I think a good place to start would be when we ask,
Q: Will agentic AI redefine the insurance core? What is the real question actually here, Mradul, insurance leaders should be asking?
Super. Let's get started. And I think a good place to start would be when we ask,
Q: Will agentic AI redefine the insurance core? What is the real question actually here, Mradul, insurance leaders should be asking?
Speaker – Mradul - 02:06
Yeah, I mean that this is this is an interesting place, and we are at an interesting point right now with a lot happening with AI.
What I believe is like the industry is not asking the right question. Okay, I feel the debate should not be whether agentic AI make the core is changing the core or is making the insurance core platform obsolete, but whether it will fundamentally reshape insurers' business architecture, operating model, and how the work is getting done on the ground. I have worked with lot of insurance companies. I know like any of a core modernization requires 10s of millions of dollar, and so it is very natural for executives to ask question whether those investment remains relevant in an AI-first future.
My view is the core systems are not going away, but they are evolving. What do I mean by that? Is the conversation is shifting from your traditional system modernization to your operating model transformation. The key question is how AI agents, human, and core platform will work together in future to deliver better outcome. And in future, the competitive advantage will come not only from core platform. Again, core platform are here to stay. They are the core of your business. How your policy gets rated, quoted, bind that that will stay. But the important thing is how effectively an insurer can orchestrate AI agent, people, and the core system across the value chain. That is the real question, and that is what is going to redefine and shape the future.
Yeah, I mean that this is this is an interesting place, and we are at an interesting point right now with a lot happening with AI.
What I believe is like the industry is not asking the right question. Okay, I feel the debate should not be whether agentic AI make the core is changing the core or is making the insurance core platform obsolete, but whether it will fundamentally reshape insurers' business architecture, operating model, and how the work is getting done on the ground. I have worked with lot of insurance companies. I know like any of a core modernization requires 10s of millions of dollar, and so it is very natural for executives to ask question whether those investment remains relevant in an AI-first future.
My view is the core systems are not going away, but they are evolving. What do I mean by that? Is the conversation is shifting from your traditional system modernization to your operating model transformation. The key question is how AI agents, human, and core platform will work together in future to deliver better outcome. And in future, the competitive advantage will come not only from core platform. Again, core platform are here to stay. They are the core of your business. How your policy gets rated, quoted, bind that that will stay. But the important thing is how effectively an insurer can orchestrate AI agent, people, and the core system across the value chain. That is the real question, and that is what is going to redefine and shape the future.
Speaker – Neerja - 03:44
Yeah, and I think Mradul, another important question. Then here is based on what you said. Core platforms have traditionally been at the center of insurance operations, and you know that may remain the same.
Q: But how does agentic AI change the role that these platforms play.
Yeah, and I think Mradul, another important question. Then here is based on what you said. Core platforms have traditionally been at the center of insurance operations, and you know that may remain the same.
Q: But how does agentic AI change the role that these platforms play.
Speaker – Mradul - 04:02
Hmm. Things are going to change again, and change is again law of nature, right?
But with agentic coming in, the power of the technology is so much that it is fundamentally going to change how the works get done, right? If I have to ask insurer to go outside insurance industry and look into what world is doing, right? What is happening is traditionally what a user will do. User, you and I, we like what we do, right? We log in into the system, we click, click, click, we navigate to the system, and we perform our daily task.
That happens in the insurance world as well. If I take an example of an underwriter today, underwriter what they do is in the morning they come. They often open multiple systems. Okay, they gather information, analyze risk, and make decisions on the policy. Even if some of the insurers have adopted underwriting workbench or a similar platform, still underwriters spend a significant amount of work in triage submission, reviewing data, conducting analysis, and making decision, and finally decide whether a particular submission is right for their appetite or not.
With agentic AI coming in, what will happen is an underwriter, and now we are focused on underwriter role, but again it can apply across insurance value chain. Underwriter will be supported by a digital assistant. What do I mean by that? The underlying technology will be agentic AI or or AI ML algorithm and whatnot, right? But an underwriter will be supported by a digital assistant. In a typical workflow, all the submissions will come. Your agentic AI will help in triaging all the submission, everything will come at one place in front of the underwriter. Underwriter will have most of his routine tasks done taken care by agentic AI, and their bandwidth will be free from doing mundane tasks. And they will focus on value and what they have learned from the business and take decision on right business that they want to take it ahead.
So, in nutshell, what will happen is instead of a user or persona moving from screen to screen, AI agent will help them gather information, execute tasks, and coordinate workflows seamlessly across multiple systems. Humans will remain in control, provide oversight, validation, and judgment and critical decisions, but they will be supported by again, as I mentioned, digital assistant or agentic. If I if I have to go into a technology term, what is happening is core system is evolving from center of work to becoming the foundation that will support the AI first future.
Hmm. Things are going to change again, and change is again law of nature, right?
But with agentic coming in, the power of the technology is so much that it is fundamentally going to change how the works get done, right? If I have to ask insurer to go outside insurance industry and look into what world is doing, right? What is happening is traditionally what a user will do. User, you and I, we like what we do, right? We log in into the system, we click, click, click, we navigate to the system, and we perform our daily task.
That happens in the insurance world as well. If I take an example of an underwriter today, underwriter what they do is in the morning they come. They often open multiple systems. Okay, they gather information, analyze risk, and make decisions on the policy. Even if some of the insurers have adopted underwriting workbench or a similar platform, still underwriters spend a significant amount of work in triage submission, reviewing data, conducting analysis, and making decision, and finally decide whether a particular submission is right for their appetite or not.
With agentic AI coming in, what will happen is an underwriter, and now we are focused on underwriter role, but again it can apply across insurance value chain. Underwriter will be supported by a digital assistant. What do I mean by that? The underlying technology will be agentic AI or or AI ML algorithm and whatnot, right? But an underwriter will be supported by a digital assistant. In a typical workflow, all the submissions will come. Your agentic AI will help in triaging all the submission, everything will come at one place in front of the underwriter. Underwriter will have most of his routine tasks done taken care by agentic AI, and their bandwidth will be free from doing mundane tasks. And they will focus on value and what they have learned from the business and take decision on right business that they want to take it ahead.
So, in nutshell, what will happen is instead of a user or persona moving from screen to screen, AI agent will help them gather information, execute tasks, and coordinate workflows seamlessly across multiple systems. Humans will remain in control, provide oversight, validation, and judgment and critical decisions, but they will be supported by again, as I mentioned, digital assistant or agentic. If I if I have to go into a technology term, what is happening is core system is evolving from center of work to becoming the foundation that will support the AI first future.
Speaker – Neerja - 06:40
Right, interesting, Mradul. With your help, we're gonna dig a little deeper here and get to the core of core.
Q: Where do you see the most meaningful value from agentic AI? Is it inside the core? Is it around the core? Or in the way work gets orchestrated across the enterprise?
Right, interesting, Mradul. With your help, we're gonna dig a little deeper here and get to the core of core.
Q: Where do you see the most meaningful value from agentic AI? Is it inside the core? Is it around the core? Or in the way work gets orchestrated across the enterprise?
Speaker – Mradul - 07:00
Yeah, I mean, I I go straight into the answer for this. As per my experience, core is not getting replaced. It's it's the heart of how insurance again transactions is taken care of. But the greatest value, what I think, is in agentic AI, and the real opportunity is lies into orchestrating work across people, system, and AI agents.
Insurance should focus less on how do I add AI to the core, and they should focus more on how should the workflow between human, AI agents and their enterprise system, and make sure that each of them play to their strength.
So organization that gets this right, definitely we will achieve lower operating costs, faster speed to market, improve customer experience, and obviously stronger regulatory compliance. And this all can be translated into insurance KPI like combined ratio, loss ratio, expense ratio, and whatnot. Right? This is all where money is made in insurance. So before we even think of scaling AI, right? My suggestion and what I have seen in my experiences, there has to be good amount of energy spent in figuring out what is our right workflow in an AI first future. And there are a lot of fundamental questions that we have to collectively again, the insurers have to answer as well whether they have the right technology, right operating system whether the system of record is actually the trusted source of truth because what happens in insurance things are scattered how we can get everything together and then feed it to AI as well.
So sorry for giving a long-winded answer, but I feel the real opportunity lies into how you design or redesign your operating model, where you have your people, your systems, and AI agent work to their best potential.
Yeah, I mean, I I go straight into the answer for this. As per my experience, core is not getting replaced. It's it's the heart of how insurance again transactions is taken care of. But the greatest value, what I think, is in agentic AI, and the real opportunity is lies into orchestrating work across people, system, and AI agents.
Insurance should focus less on how do I add AI to the core, and they should focus more on how should the workflow between human, AI agents and their enterprise system, and make sure that each of them play to their strength.
So organization that gets this right, definitely we will achieve lower operating costs, faster speed to market, improve customer experience, and obviously stronger regulatory compliance. And this all can be translated into insurance KPI like combined ratio, loss ratio, expense ratio, and whatnot. Right? This is all where money is made in insurance. So before we even think of scaling AI, right? My suggestion and what I have seen in my experiences, there has to be good amount of energy spent in figuring out what is our right workflow in an AI first future. And there are a lot of fundamental questions that we have to collectively again, the insurers have to answer as well whether they have the right technology, right operating system whether the system of record is actually the trusted source of truth because what happens in insurance things are scattered how we can get everything together and then feed it to AI as well.
So sorry for giving a long-winded answer, but I feel the real opportunity lies into how you design or redesign your operating model, where you have your people, your systems, and AI agent work to their best potential.
Speaker – Neerja - 08:56
No, that's a great answer, and thank you for going into detail, Mradul. So given this scenario, right, as platforms like Guidewire, Duck Creek, and Majesco embed more AI copilots and assistants,
Q: How should insurers balance vendor-led AI with their own enterprise orchestration strategy?
No, that's a great answer, and thank you for going into detail, Mradul. So given this scenario, right, as platforms like Guidewire, Duck Creek, and Majesco embed more AI copilots and assistants,
Q: How should insurers balance vendor-led AI with their own enterprise orchestration strategy?
Speaker – Mradul - 09:16
Yeah, I mean that's a pretty good question, Neerja, and let me give you an analogy there, right? Again, and this is my personal experience, right? When I go with my daughter, if you go to an ice cream shop, she'll leave the home and she'll say, "I want chocolate ice cream. But when she reach the ice cream shop, she'll say, "No, I want vanilla. No, I want strawberry. So, that that is what is happening in the industry. I go to Duck Creek. I go to Guidewire. Everyone will have their say form of AI. What do you say capability? And at the nutshell, the core is same, right? Whether I am doing agenting, whether I am doing a co-pilot, right? Whatever it is, so they everyone is bringing, everyone wants to show their capability.
But my advice to insurer is how we can take. Strategic approach, not just leveraging vendor provided AI where it adds value, but we'll have to maintain an enterprise level orchestration layer that enables integration, governance, and cross-platform workflow. What do I mean by that? I know my kid; she likes chocolate ice cream, but when she go out, she has variety of options as well. Can I make my ice cream inside my home? I mean, the answer can be yes or no as well. That is exactly where what I suggest insurer is. The goal is not to choose between a vendor AI or an enterprise AI, but is finding the right balance.
There are sometimes you can you want your kid to be healthy. You can make ice cream inside the home, but sometimes you want them to explore as well. Again, it is an analogy, but the crux here is how we can think of an enterprise-wide view and make AI available across the enterprise or across the insurance value chain. That is critical instead of getting stuck into a particular vendor or the industry term is vendor locking as well. That is very important, and I'll tell you a good example.
Recently, travelers have introduced again. Travelers is an insurance company. They have introduced an in-house specialized LLM tailored on their PNC business. This is a good example. While they continue to leverage technology from their platform partner, they are also building AI capability aligned to their broader business goal. That combination serve as a strong foundation for an enterprise-wide strategy, and that leads a good foundation to an AI strategy, and that can help them to integrate whether they want to pick vendor AI or let's assume they have features from Enter AI, but under the hood they can stitch in their LLM. So that flexibility and that control is my suggestion where insurers should head to.
Yeah, I mean that's a pretty good question, Neerja, and let me give you an analogy there, right? Again, and this is my personal experience, right? When I go with my daughter, if you go to an ice cream shop, she'll leave the home and she'll say, "I want chocolate ice cream. But when she reach the ice cream shop, she'll say, "No, I want vanilla. No, I want strawberry. So, that that is what is happening in the industry. I go to Duck Creek. I go to Guidewire. Everyone will have their say form of AI. What do you say capability? And at the nutshell, the core is same, right? Whether I am doing agenting, whether I am doing a co-pilot, right? Whatever it is, so they everyone is bringing, everyone wants to show their capability.
But my advice to insurer is how we can take. Strategic approach, not just leveraging vendor provided AI where it adds value, but we'll have to maintain an enterprise level orchestration layer that enables integration, governance, and cross-platform workflow. What do I mean by that? I know my kid; she likes chocolate ice cream, but when she go out, she has variety of options as well. Can I make my ice cream inside my home? I mean, the answer can be yes or no as well. That is exactly where what I suggest insurer is. The goal is not to choose between a vendor AI or an enterprise AI, but is finding the right balance.
There are sometimes you can you want your kid to be healthy. You can make ice cream inside the home, but sometimes you want them to explore as well. Again, it is an analogy, but the crux here is how we can think of an enterprise-wide view and make AI available across the enterprise or across the insurance value chain. That is critical instead of getting stuck into a particular vendor or the industry term is vendor locking as well. That is very important, and I'll tell you a good example.
Recently, travelers have introduced again. Travelers is an insurance company. They have introduced an in-house specialized LLM tailored on their PNC business. This is a good example. While they continue to leverage technology from their platform partner, they are also building AI capability aligned to their broader business goal. That combination serve as a strong foundation for an enterprise-wide strategy, and that leads a good foundation to an AI strategy, and that can help them to integrate whether they want to pick vendor AI or let's assume they have features from Enter AI, but under the hood they can stitch in their LLM. So that flexibility and that control is my suggestion where insurers should head to.
Speaker – Neerja - 11:50
Right, that's a great example, Mradul. And to sum it up in ice cream terms, don't have an appetite only for one single flavor. Right, mix it up and stay flexible.
Q: Mradul, for insurers still early in this journey, what are the toughest questions leadership teams should be asking before they scale agentic AI?
Right, that's a great example, Mradul. And to sum it up in ice cream terms, don't have an appetite only for one single flavor. Right, mix it up and stay flexible.
Q: Mradul, for insurers still early in this journey, what are the toughest questions leadership teams should be asking before they scale agentic AI?
Speaker – Mradul - 12:11
Yep, I think we discuss right what happens with world of agentic. It is pretty easy to shape up something and see something which is working, but will it work at an enterprise level? That is the critical question, right? I can create couple of agents. There are a lot of NATN and whatnot is available in the market, but is it scalable to enterprise? That is the question.
So my suggestion to insurer executives is, before scaling agentic AI, there are at broad level four critical questions that they should ask.
First one is: Are we designing work or letting technology define our operating model? The operating model, in my opinion, should come first, and then then the technology.
Second is this very technical, but this is very helpful. Can your AI interact with your core system through APIs, or does the process still rely on heavy human intervention? This is what we discuss. Where web is changing, their traditional web user is human, but in future AI will be the new web user. So there is a principle called as Mac, which is nothing but how your technology landscape is built on whether it is microservices, API first, cloud native, and headless. So that is that principle supports on ground your scaling of agentic AI.
Third question, which is very true to insurance world, is: Is your system of record truly system of truth, meaning the data scattered in insurance. How we can get the data together, and whatever we are feeding to your agentic AI or your AI is the correct system of truth.
And last one is what is your long-term AI strategy, right? Like again, going back to our ice cream example, whether it is vendor provided AI, enterprise wide orchestration, or a deliberate blend of both. So I think all of this question shift the conversation from AI hype to organizational readiness and help ensure AI scales on a strong operational foundation. And if you have a strong foundation, I think scaling up will give you good outcomes and will give you right outcomes as well. But important thing is to put the right foundation.
Yep, I think we discuss right what happens with world of agentic. It is pretty easy to shape up something and see something which is working, but will it work at an enterprise level? That is the critical question, right? I can create couple of agents. There are a lot of NATN and whatnot is available in the market, but is it scalable to enterprise? That is the question.
So my suggestion to insurer executives is, before scaling agentic AI, there are at broad level four critical questions that they should ask.
First one is: Are we designing work or letting technology define our operating model? The operating model, in my opinion, should come first, and then then the technology.
Second is this very technical, but this is very helpful. Can your AI interact with your core system through APIs, or does the process still rely on heavy human intervention? This is what we discuss. Where web is changing, their traditional web user is human, but in future AI will be the new web user. So there is a principle called as Mac, which is nothing but how your technology landscape is built on whether it is microservices, API first, cloud native, and headless. So that is that principle supports on ground your scaling of agentic AI.
Third question, which is very true to insurance world, is: Is your system of record truly system of truth, meaning the data scattered in insurance. How we can get the data together, and whatever we are feeding to your agentic AI or your AI is the correct system of truth.
And last one is what is your long-term AI strategy, right? Like again, going back to our ice cream example, whether it is vendor provided AI, enterprise wide orchestration, or a deliberate blend of both. So I think all of this question shift the conversation from AI hype to organizational readiness and help ensure AI scales on a strong operational foundation. And if you have a strong foundation, I think scaling up will give you good outcomes and will give you right outcomes as well. But important thing is to put the right foundation.
Speaker – Neerja - 14:31
Yeah, some really great insights there, Mradul. And I think the takeaway is clear: agentic AI may not replace the insurance core, but it will redefine how work happens around it, and core platforms will remain systems of record. while AI-led orchestration, trusted data, and smarter operating models will drive the next wave of value, and that's something to think about and prepare for.
Thank you so much, Mradul. For joining us and sharing your perspective here.
Yeah, some really great insights there, Mradul. And I think the takeaway is clear: agentic AI may not replace the insurance core, but it will redefine how work happens around it, and core platforms will remain systems of record. while AI-led orchestration, trusted data, and smarter operating models will drive the next wave of value, and that's something to think about and prepare for.
Thank you so much, Mradul. For joining us and sharing your perspective here.
Speaker – Mradul - 15:03
Thank you, thank you for having me.
Thank you, thank you for having me.
Speaker – Neerja - 15:04
And to everyone listening in, thanks for tuning in. Stay connected here for more conversations on the technologies shaping the future of businesses. I'll catch you next time.
And to everyone listening in, thanks for tuning in. Stay connected here for more conversations on the technologies shaping the future of businesses. I'll catch you next time.
You were listening to Tech Lyceum, a podcast from Birlasoft.
Other related Tech Lyceum episodes
