SIR Podcast
SIR Podcast
Episode 6: Manu Mazumdar, Conning
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
SIR Podcast host Tony Cañas (Insurance Nerds) talks with Manu Mazumdar (Conning) about the adoption of AI in the insurance industry and its impact on workforce and workflow.
Learn more at www.sirnet.org.
And welcome to the SIR podcast, a very, very special episode of the SIR Podcast. This is your host, Tony Kenyas from Insurance Nerds. And today I am just beside myself to welcome our guest, Manu Maszoomdar, Head of Data Analytics and Insurance Technology at Conning. Manu, thank you for joining me today. How are you doing?
SPEAKER_01Tony, it's such a pleasure. Not only is it a pleasure to join you, because you and I have known each other at various conferences. You are certainly a very prominent person at any conference that I see and have the privilege of attending, but also SIR is near and dear to us, both personally to me as well as for Conning. Conning, as you know, is a major one of the sponsors of SIR, and we uh do derive a lot of benefit with our association with SIR. So I'm really excited to be able to speak with you and to our larger SIR community today.
SPEAKER_00But if I'm not mistaken, you are in Hartford.
SPEAKER_01I am, snow-covered Hartford. So as you know, we have had uh a tremendous amount of uh weather in the last couple of weeks. So yes, but I've lived in the Northeast for uh the better part of the last 40 years, so um I'm I'm a hardcore northeasterner, and so this is part for the course, you know, Boston, Connecticut. So I'm I'm from this area.
SPEAKER_00I I spent 10 years in Iowa. Uh so I'm I I'm used to the Midwest winners, not quite northeast winners. I only go to the northeast when the weather's nice. Uh but after 10 years in the uh Midwest, uh I recovered my my uh my sanity and moved south. Uh and uh I well moved around, but but I'm in Atlanta, Georgia nowadays. Uh and uh it's cloudy here, which is not great for for for winter here, but I will take this over all my friends in the Northeast who have been under like feet and feet of snow over the last week. So good luck to you guys getting through the rest of it.
SPEAKER_01We are digging ourselves out quite literally.
SPEAKER_00I spend a lot of my time uh on my other podcast on profiles and risk interviewing insured tech founders. And for the last couple of years, all they talk about is AI. For the last year and a half, all they talk about is agentic AI. And it's very hard for me because I I'm non-technical. So it's very hard for me to differentiate what's for what's for real and what's not, who is actually building their own models and who is not. Uh so I am super excited for today's conversation because uh Manu uh leads uh a survey, a yearly survey at Conning of the top 115 decision makers who approve AI spending in the insurance industry. So while he can't name them out loud, we're talking basically about CIOs and others that report directly to CEOs and sign the checks that actually fund this in insurance. So basically, this yearly survey uh we all wait for with bated breath. Uh so having you present the results of the survey here is just like I lose sleep over how exciting this is.
SPEAKER_01Thank you so very much, Tony, for that for that introduction.
SPEAKER_00So, yeah, so today's conversation completely focused on how AI is impacting the insurance industry, and the floor is yours, my friend.
SPEAKER_01Thank you so very much, um, Tony. And you know, uh you are correct uh in the uh in mentioning that I am privileged to be the lead author uh and the lead for not only just authoring the uh results, but also uh speaking with and interviewing uh the C-suite executives who uh greenlight the decisions to implement and adopt AI in various parts of their industry. Now, the survey that you mentioned, we focus it exclusively on the insurance industry. So uh these are all uh insurance industry executives. Last year we were in 2025, uh we were privileged to be able to speak with 115 uh insurance industries, uh, tippy top, C-suite, or some of them were you know, by their titles were VPs or senior VPs, but they are the decision makers who are green lighting these decisions. And you know, uh Tony, the reason why we started doing this, this was the third year that 2025 was the third year that we did the survey in a row. We are in the middle of actually uh planning for the 2026 survey. We usually uh uh release our results of the survey sometime in uh June, July format uh timeframe. And so we are literally working on that for this year as well. But this has become one of, like you said, the preeminent uh survey-based uh studies that the insurance industry's executives uh wait for and look forward to. And so why is that, right? Because technology has literally transformed the insurance industry over centuries, right? I mean, if you look at from the birth of factorial science, which forms the basis for the insurance industry, to current day artificial intelligence, AI, which everybody's talking about, the story of technology and its progress within the insurance industry is just fascinating. And you know, AI is no longer Tony and emerging technology. It is in fact driving real results across the entire spectrum of the value chain of the insurance industry. It has evolved from just a concept, just about, you know, it's kind of funny. In in the insurance industry, we are used to talking about you know evolutions in years or decades. The AI evolution we talk about in months. So just about 24 or 36 months ago, it was a concept. And then over the last 24, 12 months, we've seen that it's actually making real impact. And um, whether it's uh in underwriting or in claims adjudication, all of those parts of the value chain are seeing real results, right? Um, and you know, the US insurance industry's journey with AI reflects an ongoing balance. It's leveraging cutting-edge technology to automate and enhance decisions, but while maintaining significant human oversight and trust. And the reason for that is the insurance industry's fundamental pillar is based in trust, right? It's a trust industry, and so the insurance industry is very cognizant of that. And with AI continuing to advance, insurers, I think, are poised to further refine risk predictions, streamline claims even more than what has already been done, and incorporate AI in all parts of the value chain so as to improve the overall customer experience in the years ahead. Because, like I said, it's a trust business, and the two counterparties of that trust value chain sit between the insurance company and its customers. So that's part of the reason why we're seeing uh the human in the loop, so to say, uh, at every stage. So having said that, uh, I know you've got a couple of questions, but before you ask me a few questions, I got to uh you know make sure that I talk about my paymaster, Conning. Um, and so Conning, for uh those of you who are not familiar, Conning is an insurance asset management company. Uh we have been uh we we were founded in 1912, and we are uh we invest in institute, we are an investing organization for institutional assets for over those decades. Within Conning, a major differentiator is our insurance research organization, and that's where we have the association with SIR, which is also a conglomeration of research uh entities. And um with the insurance research division, we've been doing research for over 55 years, so we are deep into it. We cover 30 different lines of insurance um uh sectors, and uh we go from PNC to life to health to annuities, and of course, now technology sits as an umbrella ubiquitous across all of those lines. So um that's what conning does. Now, I know you wanted to ask some questions about the uh survey per se.
SPEAKER_00So what what what exactly do you do you do you guys look look at when when you are talking about AI in the insurance industry?
SPEAKER_01Great question, Tony. So so it you know, we want to get a viewpoint of how it's being adopted on two bases primarily, right? First basis is how is AI impacting the workflow in the insurance industry? And what is that adoption of AI into the insurance industry's value chain doing to the workforce of the future, right? And so how it's impacting the people and how it's improving the processes, uh, et cetera. And so what we've done 2026 is going to be the fourth year in a row that uh we will be doing this uh survey-based study. It's a study that we've released uh for the last three years regularly around, like I said, uh June, July time frame. So stay tuned for 2026 study that's coming out soon. But in order to get a deep perspective of how the insurance industry executives are thinking and adopting AI in their business workflow, and like I said, how that's impacting their workforce, we at Conning's Insurance Research conduct a survey of the top decision makers of the insurance industry. So, similar to prior year surveys, this year also, we are soliciting responses from, like I said, the C-suite. So, what does that mean? The CTOs, chief technology officers, the CIOs, chief information officers, chief AI officers. That's becoming a prominent role in the insurance industry now. And uh decision makers at that level who green light the adoption of these emerging technologies. So uh this year will mark the fourth edition of our survey, and we are really, really excited to share with you some of the information that we've had from the last three year survey, especially from last year's survey, and therefore what to expect for the 2026 survey. So I know you wanted to uh prior to the recording, you uh wanted to talk a little bit about uh what exactly do we look at, right? I mean, is that one of the questions that you had? Absolutely. Okay, so um, you know, when we look at AI adoption in the insurance industry, we look at the value chain. So before I deep dive into the adoption results, let me take a quick minute to talk about the components that we measure and how we measure each of those, right? So we considered four components of the value chain: sales and underwriting, operations and claims, risk control and pricing, and shared services, which incorporates HR, IT, etc. Those are the four components of the value chain that we look at to see how and where AI has been adopted. We ask respondents to rate their adoption for each part of that value chain component and for each of the AI technologies, for example, generative AI, which also now includes large language models, the machine learning and predictive analytics, speech recognition, or any other technologies that they are adopting within the AI component. And for each AI technology, we ask respondents for the stage of adoption. What do I mean by that? We go from asking them, are you conceptualizing this? Meaning we are thinking about it, but you haven't really done anything with it. Are you piloting it in very select parts of it? Are you in early adoption? So some parts of the AI value chain in the AI value chain, you've adopted this, or have you fully adopted this and are achieving meaningful results of that? So that's what uh we look at as a result as part of a value chain, but also what specifically do we ask relative to each component of the value chain? Does that answer your question?
SPEAKER_00Absolutely. Um so what are what are the results this year?
SPEAKER_01Okay, so before I tell you about the results, I've got to tease you a little bit. Uh I I have to talk about who did we survey, right? So I talked about that we surveyed 115 C-suite executives, but what were what were the parts of that? So let me give you some more detail on that. So 44% of the respondents came from life and annuity insurers. An equal number, 44% of the respondents came from PNC insurers. 10% of the respondents, they came from multi-line insurers, right? And 2% of the respondents, we deliberately include that group, are companies that uh provide services to the insurance industry exclusively, so that we get a well-rounded perspective. So 44% each, 44% and 44 from PC and Life, 10% from multiline, and then 2% from service providers exclusively to the insurance industry. That's what uh we looked at. And so within that, you asked about what were some of the results. So, drum roll, let me tell you some of the results. So, interestingly, 34% of the respondents they indicated that they have fully adopted AI as part of their value chain. Now, 34% doesn't sound like a lot, but let me give you a comparative. 34% in our 2025 survey results, right? That came out uh towards uh the middle of last year. That is a 400% increase from our prior year survey, right? So that's a major change. Second uh key findings incorporated 89% of the respondents, they indicated they are in some stage of adopting AI as part of their value chain. So 34 fully adopted, 89% said they're in some stage of adoption. Now that is a 12 percentage point increase from the prior year survey. So that shows you the rapidity with which AI is being adopted in the insurance value chain. One more interesting factoid 90% of the respondents are in some stage of implementing generative AI. Now, to think about that for a second, the prior year survey, generative AI literally didn't even show up on the scale. So in 20 in about 12 months or so, generative AI went from zero to 90% of the respondents saying they're in some stage of implementing. More interestingly, 55% within that component, they said they're either in early or full adoption. So oftentimes I say that generative AI in the insurance industry went from zero to 55 in just 12 months, right? That's incredible. And then the other part that's very interesting is what we found from the survey was that machine learning and predictive analytics that has almost permeated across the board in every part of the value chain. And there was 74% full adoption, utilization of machine learning and predictive analytics across the entire part of the value chain, which in my opinion seems like that is something when we're going to release the 2026 survey. Does that answer your question?
SPEAKER_00Absolutely. Absolutely. All right. Um where where are the trends moving when it comes to AI?
SPEAKER_01Very interesting. So year over year, let me give you some um comparatives as well, or uh more comparatives. So the significant uh advantage of us conducting the survey year over year, of course, is that we have the ability then to compare results year over year, right? And so the swift progression with within a single year is evident from the increase in full adoption rates of AI. So, to give you some numbers around that, for large language models, we saw an increase from 3% full adoption in the prior year to 18% full adoption in this last year's survey. Machine learning predictive analytics, like I said, full adoption, right? Went all the way from you know about 47% to about 74%. Speech recognition grew from 8% to 40% full adoption. Now, this trend um can be better understood when you examine the percentage of technology in the early adoption or pilot from year to year, right? That's why I wanted to share those uh numbers. So additionally, if you think about it, there the the movement of all of these uh technologies becoming prominent parts of service delivery within the value chain indicates that a strong likelihood of widespread adoption as insurers continue to explore and leverage applicable use of generative AI. So we are very, very excited to see what the results will be for the 2026 area.
SPEAKER_00Did we already talk about workflow and workforce changes that you're expecting?
SPEAKER_01Yeah, so we we talked a little bit about that. So, you know, uh oftentimes when I'm speaking at various conferences, Tony, I'm asked, so Manu, you've shared a lot of information around how AI is being implemented in the workflow of the insurance industry. Can you talk a little bit about how that's impacting the workforce, right? So huge, right? So in a way. What I uh oftentimes start out with is by saying that you know there's been a significant growth, right, there in uh AI-driven um uh jobs. So oftentimes people kind of say that, oh, will AI you know kind of uh take my job away, right? There is this uh nascent uh but real fear amongst the workforce around how is AI going to upend everybody's uh jobs and careers. And you know, there are certain jobs that will be changed, uh maybe forever, but for each of those jobs that are changed forever or diminished forever, there will be other hundreds of thousands of jobs that will be created on the other end. So, how do I come to that conclusion? Let me uh give you some statistics. So we looked at the Bureau of Labor Statistics data that they come out with, right? And when I when we looked at that and analyzed that over a long period, so from 2002 to 2024. Now, you know, BLS comes up with their data a year lagging, right? So that's the latest number of uh data years that we'd had. But if you look at that 22-year period, what you see is that there has been a significant growth in higher paid and higher skilled occupational categories in the insurance sector. What do I mean by that? What I mean by that is if you look at the computer and mathematical occupation vertical that the BLS uh publishes, right? That's a category that they publish in their data, which includes data science, actuarial, and similar roles, those roles have increased by over 70% from 2002 to 2024 in the insurance sector. And with these occupations averaging about over$110,000 per year in wages, right? Managerial positions have also risen by over 70% with that set of data that we looked at, with average annual wages of nearly$180,000 per year. Additionally, mid-range occupations, such as those in sales, business operations, they have also shown considerable growth over the past two decades. Now, there's uncertainty about whether growth will continue in the way AI is literally reshaping as we speak the workforce, or if these career paths will also be disrupted. But historical examples, including typing pools and human calculators being replaced by word processors and spreadsheets, have shown that each of those categories of jobs that get eliminated, the future jobs that get created are much more satisfying to the human workforce and are much more higher paying. So if that trend over each of the past industrial revolutions continue, AI is oftentimes referred to as the fourth industrial revolution. We uh, you know, I believe that that trend is uh gonna hold true.
SPEAKER_00So, Manu, if if if I may, this this is fascinating and it makes me feel a lot better. So I get that question all the time in chatwithonia.com, which is my free career advice service. Um the it seems to me that all those prior transitions, uh basically the key for the employee, the key for for the professional has been just get more education and you you can rise, uh, or maybe your kids can rise to to a much better, a much more satisfying, much better paid role. So my fear with with AI is that we'll get to the point where more education doesn't do it because AI is smarter than I am in every field. Is is that something we're at risk at, or or am I being paranoid on that one?
SPEAKER_01Well, you know, that's a good question, Tony. And um, you know, I don't have a crystal ball in front of me um to be able to predict uh you know 100% accurately, but I can from all of the historical analytics that I've seen and for from all the research that you know you know, I am deep into AI research. You know, I um I have had the privilege of speaking at multiple AI conferences around the world. I was one of the speakers at a global AI conference in London, I was a speaker uh for a conference in Zurich. I was a speaker at the national uh conference for the National Association of Insurance Commissioners last year in December, uh, where I talked about AI's adoption in the insurance industry. And most recently, a couple of weeks ago, I was uh uh privileged to have been called by the government accountability office uh who uh wanted to understand my thoughts on what AI is doing and how its adoption and changing in the workforce. So, from all of the research that I've done and from all of the discussions that I've had with significant uh interested and intelligent minds in the industry, here is my analysis. So AI is definitely revolutionizing all parts of the workforce, but the workforce comes from the educated folks, right? So, education to your question is also evolving, right? What do you study? How do you study? I'll give you an example, a personal example, Tony, which will resonate with you and answers part of your question. So I have one child, one daughter, right? And uh she, of course, now is 22 years old and just graduated uh from college with a double major and now is working for, lo and behold, Apple doesn't fall far from the tree. She's working for one of the top insurance companies right now in the finance division. But this story goes back a few years ago. She was uh about eight or nine years old. Uh, and on one Sunday morning, I was sitting and uh uh having a cup of coffee, and she was running around in our living room like a child would, like eight or nine-year-old child would. And uh Sunday afternoon, so I called her. I said, Come here, my dear, uh, my only child, right? So um, and I said, Hey, have you finished your homework? She goes, Yes. Have you finished the extra reading? Yes. Have you finished the extra math? Yes. All of that stuff, right? Yes, yes, yes. And then, so, and then she starts playing again. So I said, Wait, I'm gonna ask you a question. So, off the top of my head, I ask her, hey, what's the distance between the earth and the sun? Sun, right? So she, this is her answer. A nine-year-old at that time gives me the following answer. She says, Papa, in your day, you would need to learn information like that and remember information like that. Now we have Google to search for information like that that we don't need to retain, but what we need to do is to understand how do we apply that searched knowledge that we get from Google. This is a nine-year-old, remember, right? And so I'm looking at her, and so then she goes, huh, again, puffs, turns around and starts walking away three steps into it, turns back to me and says, but in case you're wondering, it's 93.1 million miles, right? Right? So, but point is right, that in information and how it's digested is very, very different from generation to generation. Educational institutions and educational patterns are changing. And I'll give you a practical example today. You know, you asked about education and how the workforce is being impacted. So let me draw your attention to what we've done at Conning. We've done a 10-year look of what skills are important today in the insurance industry and what skills would be very important for the insurance industry 10 years out. Now, one of the prominent skills that any industry, of course, the insurance industry needs, is computer programming, right? You need programmers who program and uh do all of the applications that's needed for us to do the businesses the way we want to do. Now, with the advent of AI and various agentic AI abilities and tools, AGI, AI agents, you mentioned at the top of the conversation, when you think about what that is doing, programming skills are becoming less and less and less important. And what's replacing that is prompt engineering. So, what I what I oftentimes say at conferences and speeches is look, there it was a P word-related position that was extremely important, and there is another P-word-related position that is being replaced, is replacing that other one, programming to prompting, right? Better prompting is delivering much more results because AI is delivering what in the past you would have to have hardcore programming for. So, education, skills, all of those are changing. And you know, in addition, if you think about it, you touched on it, technological literacy and digital fluency is becoming more and more and more prominent. But the understanding of how to work with a human-machine collaboration, we've never had this in the past where you know what it is completely imaginable and is going to probably be a reality in the next 24 months or so, if not currently already, where you're gonna have an AI agent as one of your colleagues sitting next to you working, right? And so that sitting is uh a strong way to put it. Yeah, so human and machine collaboration and understanding that hybrid workforce management is gonna become very important for managerial skills, right? Uh, critical thinking and problem solving, because you know, there was a time when it was routine work. Now it's becoming because the customers are demanding more, their time frames of response responses from insurance companies is truncated. So critical thinking and problem solving is going to become a very important part of the education of the new workforce. And then creativity and innovation will always, always remain in the forefront as we foray into this whole new world of having imagine an AI Asian sitting as your next uh you know, colleague in your next cube or next uh desk, right? So those kinds of things. And lastly, I think Tony, one of the more critical educational components is uh going to be a combination of emotional and social intelligence, the ethical judgment and decision making, and regulatory understanding of this. But I want to end this question of yours by saying that we have been a race where humans have adopted lifelong learning. I think that continues. Lifelong learning and adaptability to our learning is going to continue to be critical. I shared my daughter's story with you. Same thing. We will all have to adapt in terms of how our learnings are evolving.
SPEAKER_00Manu, thank you so much. That was fantastic. It was downward beautiful. I lost love the story about your daughter. Uh personally, uh, I I grew up in the everybody should learn to code world. Uh and uh as a nerd and uh a gamer and and you know, somebody who's been playing on a computer since I was very, very young, uh, I started college uh for computer science. And then it turns out that I'm really bad at math. Uh and if you if if if if if if after six attempts you can't pass Calp 1, uh you will not graduate from computer science. So I ended up graduating with a management information systems degree uh with some regrets about not having been able to follow the computer science world. Uh and and I've remained a nerd. Uh and my career has been amazing. But looking looking, I don't have kids. But if I had kids today, uh my my dad and my grandpa spent a lot of time because on one side of the family, on my dad's side, it's engineers and entrepreneurs, on my mom's side, it's all doctors. I'm the black sheep. Uh very well, Tony. Uh, so so uh but if if I don't have kids, but if I did, what would I advise them? And I think what I will advise my my nieces and nephews is uh liberal arts. Because what you're talking about, the the uh you know ethics, uh problem solving, uh being able to direct those superpowers that AI is giving us. You no longer have to be a developer to to have those superpowers, right? A little prompt engineering, which which is a lot easier that than being a developer, can give you superpowers. Uh, but how do you use those superpower powers wisely? That is the definition of a liberal arts degree.
SPEAKER_01You know, Tony, you're so correct in saying that. I mean, it's it if you think about it, throughout the history of the world, right, we've always kind of depended on the philosophers of the world to have given the thoughts and vision of what the evolution of the human race is going to be. What you just talked about is no different from what we've seen for centuries and centuries and centuries. But that trend, irrespective of what the particular focus du jour might be, that overall overarching trend around philosophy, life lessons, liberal arts, like you said, understanding of culture, history, nuances, understanding each other, that will that is ubiquitous and will continue to stay uh through the next several thousands of years as well.
SPEAKER_00Fantastic. We we we're almost out of time here, uh, but any final words? Uh the microphone is is yours. I I uh sure. I I I can't possibly uh close this any better than you just did.
SPEAKER_01Well, I'll give you some, you know, a couple of quick parting thoughts to bring it back to the AI conversation. Philosophy, of course, is very, very important, and I I respect that. But if you think about it, so a couple of parting thoughts about AI. I think AI is creating efficiencies, right? From internal chatbots to AI assistance and AGI for sales to the generative AI utilization in marketing content, right? Um AI is driving efficiencies, it's reducing costs and freeing up human capacity for much more higher value work. So AI certainly is creating efficiencies and it's starting, it's here to stay, right? And continue. The second parting thought I'd I'd I'd uh give is the rise of the AI literate workforce. We were just talking about that, right? Prompt engineering, the ability to become fluent in uh understanding and generating responses out of AI. AI fluency is becoming a core skill, right? The roles and responsibilities of the workforce will continue to evolve with AI, and the skills to use it effectively is going to become more and more and more important. You you use the uh example of prompting. So you know, prompting is also better. The the better we do the prompting, the better we are going to get the results out of it, right? So that is a core skill that I think is gonna uh continue to evolve. I think customer experience, that's the third parting uh you know, kind of uh thought that I have customer experience is getting smarter. And what do I mean by that? So AI is enabling hyperpersonalization, right? You know, I'll give you one practical example. There was a time when you know Amazon delivered in two days and I was very, very happy, right? Now Amazon delivers in four hours, and I ask, why not in an hour, right? And so where's my drone? Exactly. So customer experiences are changing and it's getting smarter and it's becoming hyper-personalized. In the insurance industry, it's becoming hyper-personalized when I walk up to um uh you know, and call my insurance agent, I expect, or call my insurance company, I expect them to already know who I am when the call comes in, you know, all of those things. So it's becoming very personalized. Uh, and that's going to continue. Data-driven journeys are helping make that personalization possible and therefore the customer experience better. However, the human touch, especially around customer empathy, must continue, right? We talked about that. So that's very, very important. And I don't know of um robots that comes with empathy built in, at least now. Who knows? Down the road, we might be able to program that in as well. Don't know. But customer empathy, customer experience, smarter customers, uh, that is a component that I think is going to evolve and very important in the AI world. And last but not least, responsible adoption is absolutely key, right? The speed to adoption should not be at the expense of a thoughtful change. It is critical, in my opinion, to balance innovation with empathy, governance, and continuation and solidification of trust. Because at the end of the day, the insurance business fundamentally is a business of trust.
SPEAKER_00It uh Manu, it has been a privilege to uh to host you today. Uh and uh I can't wait to run into you on the road sometime this year. Uh thank you so much for everything today.
SPEAKER_01Tony, yeah, uh I have to thank you for being a fantastic host. I appreciate that. You your your affability at every conference is just absolutely uh energizing. And uh I can't wait to spot the man with the red suit and the red hat and the red glasses at various conferences that I I I am quite easy to find.
SPEAKER_00Have a magical weekend, my friend. You too.
SPEAKER_01Thank you so very much for the conversation.