We explore the artificial intelligence trends and innovations that could define the insurance industry in the years ahead, and what they mean in practice for businesses and brokers. Our new host, financial broadcaster and journalist Georgie Frost, is joined by Akhil Lalwani, Chief Data Officer at Allianz UK and Mary Bowie, Managing Regional Counsel UK at Verisk, a global insurance, data analytics and software solutions provider. In this episode, we discuss: How AI is changing insurance today and what that means for brokers New exposures, accountability and the importance of data stewardship and governance Evolving customer behaviour, regulation and how to ensure AI-driven decisions are fair, explainable and accountable What the future of insurance could look like as AI matures, and the role brokers and human judgement will continue to play Follow the series for notifications about future episodes. You can watch the podcast on Spotify and YouTube | https://www.youtube.com/playlist?list=PLoLD_-KfE3a3mR38xel2jA58gVo3J6-Gs Want to help shape our future episodes? Share your ideas, thoughts and comments here | https://info.allianz.co.uk/insurance-tomorrow-podcast-survey.html Connect with us on LinkedIn [https://www.linkedin.com/showcase/allianzukbroker] for broker insights and explore the Allianz Knowledge Centre [https://www.allianz.co.uk/broker/knowledge-centre.html] for webinars, articles and resources on business risks and other key topics. A podcast series by Allianz Insurance, produced by Fresh Air Production.
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
Georgie FrostHi, I'm Georgie Frost, and I'm excited to be joining the team as your new host of Insurance Tomorrow, brought to you by Allianz. I'm a financial broadcaster and journalist who's worked with some of the UK's biggest media organizations. If you're listening for the first time, I'll be exploring the trends, technologies, and risks that could define the years ahead. The challenges your clients are facing are changing fast. So we're here to unpack some of the biggest issues. We'd love your help to make this series the best it can be. Is there a topic you'd like our experts to discuss? Get in touch via the address in the show notes. One of the hottest topics facing the insurance
Georgie Frostsector right now is artificial intelligence, highlighted by its position near the top of Allianz's 2026 risk barometer.
Akhil LalwaniSo it's all about how do we keep our customers safe? How can we price our products better? How can we make it more competitive for our customers and also give them a great experience?
Mary BowieI think brokers of the future will need to demonstrate that they are not just skilled and adept at placing activities, but they will also need to demonstrate a level of proficiency and understanding of the tools, the technology tools, and the processes that they use.
Georgie FrostWith me today to discuss AI and the impact it's already having on brokers and businesses is Allianz UK's chief data officer, Akhil Lalwani, and Mary Bowie, managing regional Council of UK for global insurance data analytics and software solutions provider, Verisk. Back in 2018, when Insurance Tomorrow launched, AI was still an emerging technology.
Speaker 4Then you can have a related issue, dramatic failures such as flash crashes in the stock market. And that's when a single AI is remorselessly following its programming. Or when you have an unforeseen interaction between two AIs. And then I would also add cybersecurity. AI is now controlling or will control maybe autonomous vehicles. It may control aspects of our important infrastructure, and you've got a risk of really catastrophic failures, which the insurance market may have to pay for.
Georgie FrostThat was 2018. So where are we? Eight years on. Mary, Akhil, welcome. Akhil, how are you already seeing AI have an impact on the insurance industry?
Akhil LalwaniI think the impact that we are seeing is quite humongous actually. And you can look at different facets of insurance right from how we engage with our customers, how we engage with our brokers, how our colleagues are using that internally to how we price risk and how fraud is evolving in the space as well. So you can look at the entire life cycle of insurance and actually visibly see the impact of AI.
Georgie FrostWould you use the word transformation yet?
Akhil LalwaniI think we are on the journey of transformation because unlike I equate the advent of AI to what we saw with the internet. And as you appreciate, everything this big will take some time to germinate and for people to accept what is actually possible with this technology. So yes, I believe it is going to be transformative. It's not yet, but it is getting there.
Georgie FrostWhere do you think it's having the biggest impact day- to- day?
Akhil LalwaniA lot of the impact that we see in the world of insurance and generally is in the operational efficiency, how we do day- to- day tasks, which are more repetitive, where we can see AI adding a lot of value and being almost an assistant to people who are actually making decisions on the frontline and the back office. A simple example is we have brilliant underwriting teams and they know the risks that they want to select. What we are now helping them do is look when they're busy, when they have hundreds of them, how do you sift
Akhil Lalwanithrough them quickly and come up with a five, 10 top risks that we know we are going to write? So take that, give them that assistance in doing what they're doing much better.
Georgie FrostMary, where are the new exposures for brokers?
Mary BowieGeorgie, I think it's quite interesting really thinking about new exposures. So we've got, in my mind, three categories of new, emerging exposures for brokers. Data exposure, so how data is for claims, for underwriting, for pricing is collected, shared and analyzed across various platforms and tools. You have the decision exposure, which is how the same categories of data
Mary Bowieare used in claims, underwriting, and decision- making outcomes. And finally, you've got governance exposure. So being able to demonstrate not just robust on paper governance, but how it is deployed throughout an organization, how it's operationalized.
Georgie FrostAnd how are brokers responding to these new exposures?
Mary BowieIn terms of response, I think it's clear that brokers are developing their own products and tools and API'ing with other tech vendors, which are keeping pace with the emerging technology that is empowered by AI. I think where gaps could emerge are in terms of definitely from a legal and regulatory and risk compliance perspective, accountability, oversight, but more
Mary Bowieimportantly, understanding how those tools behave. I think that's something that brokers need to be mindful of.
Akhil LalwaniThat's a brilliant point because I think accountability is a key question that everybody's grappling with today. And the simplest way it has been addressed so far is we always have people in the activity, no matter what AI does, so the accountability sits with people. But we are not far away from a world where there will be things that AI does independently. So where does the accountability go? Who makes the decision? What is the lineage to make the decision? The bias that this could introduce? How do you tackle and monitor that? Big questions that we have to answer as
Akhil Lalwania financial services institution because we do have the requirement and the need to keep our customers safe and our brokers safe.
Mary BowieAbsolutely. And I think it's very much, there is a regulator focused on outcomes- based decisions and making sure that those decisions are fair and transparent for clients. And interestingly as well, I think brokers have always needed to have historically been held to a high duty of care towards their clients. And I think the scope of that duty is going to develop in pace with emerging technology, meaning that brokers will need to demonstrate very active data stewardship.
Mary BowieSo to show as to my previous point, how that data stewardship is operationalized in their organizations. And that governance isn't just a policy on an intranet, it's something that is put into practice in terms of actual practical oversight. And again, to my previous point, understanding how the technology works.
Akhil LalwaniAnd this actually directly links to the disruption that we are seeing in the market around people's behavior. So look at me today. If I need an insurance policy, six months back I would go to Google and say, " Where can I find the best insurer?" I might go to a PCW's website. Today we find a lot of customers starting their journey in tools like ChatGPT and having a conversational experience to get to the outcomes that they're after. So it's very interesting. In that scenario, are we looking at ChatGPT as
Akhil Lalwanioffering financial advice? Is something we look at, financial services do. Are they regulated the same way? Do they have the same accountability that our brokers and intermediaries do? I think it's a brand new world. It's a brave world, but it's a confusing world as well. Even from a regulatory standpoint as to how this market, how far do you want to disrupt the market? Which is never a bad thing. We obviously need some disruption, but where does the accountability sit and what does that constitute at the end of the day?
Mary BowieYeah, that's a really good point. And actually thinking along those lines in terms of being an in- house lawyer and looking at how legal risk and regulatory monitoring frameworks work internally, you've got some complexities introduced by AI tools because you have large language models generate huge, large volumes of data, mixed data sets, and then decision logic that isn't always intuitive. And brokers have to learn how to
Mary Bowieinterpret that, analyze it, and add a level of human judgment that isn't just ignoring their need to be accountable and have oversight over what the AI outputs are. We've always said, " Keep a note, keep a record." And this is where internal auditable records are going to be really key to demonstrate how AI is used to influence those decisions, again, as we've said on claims, pricing, and outcomes for the client.
Georgie FrostYou mentioned they're the human touch. I want to get into that in a short while, but before we do, you also talked about AI tools. So Kil, tell me about 1BRIAN.
Akhil LalwaniSo Alliance has a tool called 1BRIAN, which is quite interesting because we have a lot of knowledge that's dispersed right across the organization. And the whole idea was to use the concept of, which started with BRIAN to service a part of the business and we've now scaled it to the entire organization in the UK and hence it's 1BRIAN where you can go and get the information that you need at the tips of your fingers. What we've essentially done is ingested all the organizational knowledge in a tool, to put it simplistically. And now you can interact with this
Akhil Lalwanitool like you would interact on a ChatGPT. So if you think of an agent on the ground who's servicing a crane, let's say, and it's of a specific make, they can just go in and say, " I'm servicing this claim, this make, this year, service history, what do I need to do now? Where can I get the service parts for this? What are the next steps?" So it just makes the interaction a lot smoother, a lot easier. Think about our underwriters and they need underwriting guidelines, which are fairly complex about what can they do, what can they not do? Now, instead of having to go through the entire
Akhil Lalwanidocument, which at times can be 30 or 40- odd pages, they can simply ask a question and say, " Am I authorized to do this? What are my limits? Am I breaching something?" And it's just there for them as a part of their day- to- day workflow, they can interact with and get the information that they're after.
Georgie FrostMary, listening to Akhil talk there, and you both, this is moving so fast. What we can do today is completely different to what we could do two years ago, and I imagine two years, hence, it will be completely different picture again. How is regulation, legislation keeping up, particularly FCA regulation? Where are the gaps?
Mary BowieBack to my core theme here of internal oversight and governance. I think brokers of the future will need to demonstrate that they are not just skilled and adept at placing activities, but they will also need to demonstrate a level of proficiency and understanding of the tools, the technology tools, and the processes that they use. So working alongside their product and tech teams and actually becoming perhaps in their own right, innovators, smart prompters, developers, and coders. I mean, we've
Mary Bowieseen that Claude and Copilot, you can do some vibe coding in there and we may even see that developing. Again, with the necessary guardrails that the regulator will expect to see. And the regulator has created a number of regulatory frameworks such as the PRA outsourcing regulation, which demands that brokers have oversight of their tech vendors. So brokers need to be mindful as well that they can't just outsource governance to tech vendors. They have those obligations to
Mary Bowietheir regulator.
Georgie FrostAkhil, Mary spoke about the brokers of the future. What muscles do they need to be building now to be prepared for that future?
Akhil LalwaniI think it's not just the brokers, but I think everybody in general needs to have a far more flexible change muscle and the ability to actually adapt. And it's at both ends of the spectrum. If you think about from an innovation perspective, there's a big question: will the products that we service, that we provide to the market today, will they remain the same? So I'll explain what I mean by that. Essentially, we are looking at an era where we might have cars that drives themselves. Now, are we now going to insure the drivers anymore because they're
Akhil Lalwaninot driving the car. My son, who's 12 now, told me the other day, " Dad, I don't think I need to learn to drive." I'm like, " What?" He's like, " Yeah, because the car will drive itself." And that got me thinking, what is the insurance product of the future for motor insurance going to look like? Am I going to insure the driver? Or am I insuring the car manufacturer because the accountability now sits with them? Is it more of a fleet insurance that we are talking about? And so we need to be ready to pivot based on where the market takes us. That's one end. On the other end,
Akhil Lalwanithe accountability question, which is a brilliant one and slightly unanswered because somebody said the other day, " When AI learns something, it always remembers." So today, we as financial services institutions have the accountability to keep that data safe and give it back to them when they ask of us. And that's a service we provide. How do we do that in the age of AI where AI has learned something and we don't know how to teach it to unlearn something? So there are a lot of open questions and that's why we need to approach ... We need to embrace this. I
Akhil Lalwanidon't think we can fight the tide, but how do we do it in a way that keeps us at the cutting edge of innovation, allows us to respond to the needs of our customers, but at the same time keep them safe from what they might not be aware of is a critical step in the entire process for adoption of AI. It's this balance that we have to strike, this pendulum that we are all trying to balance very carefully, but we haven't reached that equilibrium yet.
Mary BowieI would say actually to your point, Akhil, it's quite interesting how coverages may develop in the future, and that's where brokers can really add the value of their human judgment. You talked about your son in the next generation. I've got a nine- year- old who comes home and talks to me about discernment and deep fakes, and it's great. We're adopting that discerning mindset, which is keeping pace with the technology. And I'm a tech optimist, so I think
Mary Bowiethat's a great muscle to your point as well about a flexibility that's also strengthening.
Georgie FrostWhich leads us onto the human touch. Where does it sit in this AI world?
Mary BowieFor me, brokers really add value and really can add their human touch to the claims process. I think that's historically where brokers have always been able to be a very human connector and advocate between data under a policy and the outcome that is required for fair and principle- paced adjustment of claims.
Georgie FrostAkhil?
Akhil LalwaniSo I think we are in the business of selling trust. We provide trust to people, and that trust is most relevant when people need us. And that's when they're most vulnerable. They've had a loss, they've had a claim. Their house is on fire, their business is not working, as we saw in COVID. And that is where, for me, the human touch is very, very, very important. Because we might look at the front part of the process, " All I'm doing is buying a policy," but when you're in the need, as Mary said, when you want that human interaction, that is
Akhil Lalwaniwhen you're in distress. And I don't think that need is going to disappear. We are not in a place where we are very comfortable as a race trusting machines for the whole journey of the process. I do feel strongly there'll be parts of the process that will get distracted by what we're trying to do with AI. There will be lots of efficiency players. There will be innovation in this place as well because of AI, but there is definitely a place for us as humans to be a part of that value chain and bring people along that journey and give them the comfort that we are there for them.
Georgie FrostAkhil, I want to focus now on AI risk modeling. Just explain to me what best practice looks like.
Akhil LalwaniI can share what I believe. I think from a best practice, it might be challenging because different organizations are different places in how they model their risk. I think right now we are at a stage where there's a lot of information that we can process to better define our risk models. I think the one thing that AI does really well, depends on how you use it, is to drive differentiation because at the end of the day, we all have the same product to offer. We are insuring risk, it's car, motor, homes, buildings. It's how we assess that risk based on what we believe is the best way
Akhil Lalwanito offer a service to our customers, give them a great experience, and also lower our operational expense when you're making that decision. So if what we could do, if we could process 100 data items yesterday, now we can process 1, 000. If that would take five hours before, it's taking five minutes today. We are moving to a world where in risk modeling where we can take more instantaneous decisions, and it's not just from a risk perspective, it's from a coverage perspective. When we think also when there are CAT events that happen, when we start reserving for those
Akhil LalwaniCAT events, we can do them a lot faster. So it's all about how do we keep our customers safe? How can we price our products better? How can we make it more competitive for our customers and also give them a great experience? A good risk model for me is something that can stand the test of time that is measurable. I am confident about the output that I'm getting because ideally what we need to be able to do is stand in front of the regulator and explain the decisions that we are making. And a lot of the times the decisions that we make are unexplainable because we are using
Akhil LalwaniAI, it's decided for us. So how do you build transparency into those models? How do you make sure that what you're feeding into the model is accurate? Because while I know we are talking about AI, but what we have to remember is the output that AI is going to produce is going to be as good as the input that we given. So to get trusted outcomes from AI, you need to have trusted data. So are you feeding it the right information? Are you making sure you're not introducing bias into the conversation? And for me, a risk model has to be able to cope with the changes which are
Akhil Lalwaniexternal. So it needs to make sure that the context, your risk models are context aware, which not a lot of organizations do. So you need to put in a lot of effort around the accuracy, the measurement, the lineage, and providing context to your AI.
Georgie FrostMary, I want to look at how AI adoption is changing risk compliance and record keeping. Is it making firms more accountable? Or does it just risk greater exposure?
Mary BowieWe've been in a digital transformation space in insurance for best part of half a century, really, through from the very early development of placing platforms right through to the current kind of AI, embedding AI tools into various workflows. And I think, again, I'm going to sound like a broken record, but I keep coming back to the same theme of auditible record and documentation keeping, and showing that
Mary Bowiewe can demonstrate how data is used, how tools are deployed. So in a way, it's not so much greater exposure as it is how to use the governance guardrails, how to enhance ones that are already in place to ensure that we know how the technology is integrated into workflows and how it can be used to leverage best outcomes for clients, but also to demonstrate that with clear
Mary Bowiereference points. One area that I have a focus on is data use, definitely being able to show how data is used. And it goes now beyond the disclosures to clients under GDPR, UK GDPR, privacy law and regulation. It's not just purpose limitation or knowing where and how the data has been used, but more how it's been cared for, looked after, maintained, stewarded. There are very many other cross-
Mary Bowiefunctional teams that support that process. So you've got your information security teams, you've got your risk audit and compliance teams, all working in the background to make sure that there is appropriate verification of how data is used and how it's deployed in AI.
Akhil LalwaniI completely agree with what you've said. And there's an inherent risk in this world that we live in today because we all spoke about the human touch. And then at the same time, we also talk about teaching AI to do a lot of activities that we are doing today. We need to understand how this will evolve in the next step because if the broker of tomorrow doesn't know elements of what they do today, how are they going to train the AI future to do the things? And what happens in the world when AI doesn't work? Will they
Akhil Lalwanieven have the DNA to actually do the activities that now they've asked AI to do on their behalf? And that's true of everybody in this entire ecosystem. So I think that's something that we need to be careful about. To the point of documentation, I think we are lazy people at the end of the day. We all like the comforts of automation. I was walking into this place. I've been around here a few times, but I still followed Google Maps. I don't want to get it wrong. I want to make sure that I'm saving my time and doing the right thing. So we are inherently lazy. And automation, AI, all of these things tap
Akhil Lalwaniinto our laziness as well. So I think we need to be conscious that we don't over rely on these technologies and we still have the ability to influence the outcomes that we are getting from these technologies because there's a danger that over reliance could lead to complacency and complacency could lead to not brilliant outcomes that we all strive to give our customers.
Georgie FrostYou spoke about these tools as they get embedded into our everyday lives, and I'm exactly the same as you. I will go to the same place many times and still look at Google Maps. So I'm not actually learning where I'm going. I'm just relying on the technology. But what do you see as the wider challenges of ensuring AI- based products as they become more embedded in day- to- day life? And you spoke about cars, what are the sort of products and what are the challenges? I
Akhil LalwaniThink the biggest challenge I see is cultural in any organization, because what is the cultural acceptance of AI? And while we can sit and talk about all the brilliant things that we do, it's very important to not forget that, at the end of the day, we are a people- led organization. How do you take people on that journey? How do you make sure that they don't treat it like magic because it's almost like, " Wow, AI can do this. I can't believe it." But I don't think it's accurate. I don't think it's good enough. How do you make sure they're a part of that journey? How do we make
Akhil Lalwanisure that we are feeding it trusted information? As I said before, your foundations are right, you're able to measure the outcomes and people are able to influence the outcomes that they get from AI. There's not an overreliance on the outcomes, but it is seen as a tool that helps them make better decisions faster, rather than the decision is what the tool has given you itself.
Mary BowieI agree with that, actually. And I think also to the point we mentioned earlier about developing that quality and skill of discernment, not just my nine- year- old's detection of a deep fake image or adopted image, but I know where I've used enterprise- level AI such as Copilot or whatever in day- to- day work in BAU activities, really bringing it back to my own discernment, my own judgment about, does that sound like me? Is that a decision that
Mary BowieI think is correct? And ethics is also going to play a very, very big part in this. How do we navigate ethics amongst our ... Well, within our professional lives as we use AI? And even from a legal and regulatory framework, most organizations and businesses both within the insurance market ecosystem, third- party vendors, and the clients we serve, the policy holders, the insureds, probably every organization now has
Mary Bowiea data use policy or a data use request policy, particularly if it relates to development of a tech tool. And imagine yourself in a scenario where you feel challenged ethically by the use of data, whether that's insured clients or ... And so being able to really rely on that discernment factor and also have the courage to maintain trust by being able to operationalize whistle blowing mechanisms or other
Mary Bowieappropriate forums for exploring how that data is used. So again, it's not just name- checking it in a policy, it's actually how do we put those principles into practice and evidence and so that if we ever get a knock on the door from the regulator, any of us in any environment, we can stand by the judgment, the human judgment that we've overlaid onto use of AI.
Akhil LalwaniI couldn't agree more because as humans, what we bring is emotional intelligence and empathy. And yes, there might be a day that people have promised us that AI and machines will be able to do it, but until that time comes and we can prove that we can rely on that, I think in this world of trust that we are trying to offer people, that emotional intelligence and that empathy is so important when people are vulnerable. So we shouldn't forget the role that we play in this entire ecosystem. Yes, the world is going to change. Yes, there's a way of AI. Yes, we'll be able to
Akhil Lalwanido things faster, better, maybe even do things that we were not able to do yesterday. But that human touch will not disappear.
Georgie FrostYou mentioned there a lot of the things that could happen. And actually, I do feel like our conversation has focused on a lot of the challenges, and inevitably there are challenges with this new emerging technology, or the fast changing pace of this technology. But let's focus on the opportunities, the positives. What are you most excited for? The opportunities of what it could bring the insurance sector?
Akhil LalwaniWow. So I think it has the potential to truly disrupt how we treat insurance today. We spoke about some of the products that might not be relevant, but there are other products that could come into the fray. How we assess risk could be completely different. We could, instead of customers having a long wait time when they're trying to call, getting some information, asking for claims, there would be more digital channels and much better ways of interaction. As a customer, I could get more customized products since I go
Akhil Lalwanito a website today of an insurer or a PCW and I have to answer 35 questions before I can get a code and I have to compare. I think that conversational element that is more human, which is akin to what we're doing right now, to share ideas to get information. I might be able to customize my products based on my needs rather than going with what an insurer gives me. I think the opportunities are endless in terms of both having a much better customer experience, offering bespoke products
Akhil Lalwaniand how we service our clients in the end.
Georgie FrostMary?
Mary BowieFor me, I tell you what I'm most excited about. So after my private practice background working in insurance litigation at CMS, I worked in house in an insurance broker for 10 years, and I think brokers are great innovators. And I've said this a number of times during our session today, but it's that individual innovative skill development that I think is most exciting about AI. I've also mentioned about how
Mary Bowiewe all could become smarter prompters and developers and coders, and actually moving from that sandbox experimentation environment to fully embedded and well- governed AI deployment, it's actually an exciting journey. The brokers that are going to be able to demonstrate really good governance, practical governance, and really good data stewardship, and be able to answer the questions that clients raise
Mary Bowieabout that, will give themselves a competitive advantage in the market. And not just maintain existing client trust, but win new client trust, which is obviously great news for insurers and tech vendors alike as well. So yeah, everyone in the ecosystem benefits from an innovative mindset.
Georgie FrostThank you both so much for your time and your insights. What a fascinating conversation, and I wonder where we'll be if we revisit this topic in another eight years' time? Next time on Insurance Tomorrow, we'll discuss what you can expect from this year's annual BEBA conference. Insurance Tomorrow is brought to you by Allianz. Follow the series for notifications about all our future episodes, and you can also access previous episodes. You can now watch the series on YouTube and on Spotify. And I'd love to hear from you. Tell me about the topics that are most important to you and your clients. Is there a question you'd
Georgie Frostlove us to ask our experts? Get in touch via the link in the show notes.
UK insurers uncover more than £1billion of fraud every year. The true cost? Premium inflation, increased operational costs, and damaged customer trust. Brokers can be among the first to spot the signs of potential...
Insurance is changing fast - as AI transforms how businesses operate, new technologies reshape familiar risks, and climate change creates challenges already being felt today. But amid all that change, some fundamentals...
From conflict and trade tensions to cyber attacks, supply chain disruption and climate change, risks facing businesses are becoming increasingly interconnected. So how can brokers prepare when past events may no longer...
Autonomous vehicles are moving from science fiction to reality as legislation develops, driverless technology advances and on-the-road testing accelerates. But what happens to insurance when driverless cars start making...
How can businesses best defend their networks and day-to-day operations as cyber threats evolve and hackers make greater use of AI? Cybersecurity concerns topped the 15th annual Allianz Risk Barometer in 2026, making...