Executive Conversations

Scaling AI in Specialty Insurance: A Conversation with Markel’s Simon Wilson

“AI is the defining tool that will allow us to think like a startup again and get back to that genuine Founder’s Mentality.”

By Kyle Kinsell

First published on outubro 14, 2026
min read


Simon Wilson is putting AI at the center of how Markel Insurance competes by reimagining entire lines of business rather than optimizing them at the margins. As CEO, he is steering the company’s AI strategy across the organization, guided by a Founder’s Mentality that treats every process as one worth rebuilding from scratch rather than merely improving. Kyle Kinsell, a partner in our Financial Services and Performance Improvement practices, sat down with Wilson to discuss the future of specialty insurance, how AI is reshaping the economics of underwriting, and why he believes the winners of this shift will be decided in the next 12 to 18 months. Q: How do you envision AI transforming specialty insurance over the next several years? Wilson: AI should be viewed in the context of everything else happening in the industry and the wider economy—geopolitical risk, marine war insurance, US casualty, legal-system inflation, changing US Treasury rates. All of it affects the specialty insurance we write around the world. But amid all that noise, I think there’s a danger of missing what is a deeply important technological shift. In my 25-year career, we haven’t seen anything like it. AI represents a fundamental shift in white-collar work, equivalent to what the Industrial Revolution meant for blue-collar activity. From a strategic sense, we’ve worked with Bain a lot over the last two years, and one concept we’ve really picked up is the Founder’s Mentality. Founders are willing to do things differently from incumbent players. That Founder’s Mentality is at the core of our potential strategic advantage as a challenger in the specialty insurance market. AI is the defining tool that will allow us to think like a startup again and get back to that genuine Founder’s Mentality. It is very much at the heart of our business. Q: Can you share how you’ve approached the AI journey at Markel so far? Wilson: About three years ago, as large language models took off in our personal lives, we rolled out Microsoft Copilot to just over 5,000 Markel Insurance employees. In that early stage, there's always 5% to 10% of people who pick it up and run with it, but it doesn't have a material impact on the rest of the organization initially. Where I started to see a genuine change within the business on the back of AI implementation was when people could define a specific and important business problem that they were trying to solve and when they showed that Founder's Mentality and owned the outcome. That’s been a great realization for me. If you start with a tool in search of a job, you’ve got a problem. If you start with a problem that could benefit from the use of these tools, you start to see things in a different light. Q: What are the key business benefits you’re seeing from AI-enabled underwriting? Wilson: One example is in the group that underwrites warranties and indemnities—coverage for buyers in M&A deals if something missed in due diligence causes a financial loss. It’s a popular product: Sellers like it because it gets deals done, buyers like it because it protects them, and boards like it because they’re covered if a mistake slips through. The underwriter’s role is to insure this and price the risk. Traditionally, our underwriters—a group of buy-side lawyers—would spend about a week trawling a data room of thousands of financial and legal documents, giving maybe one in 10 a fair read. An underwriter could produce about four quotes a month, and we’d win roughly one in four. Given those economics, we tended to pursue only larger deals. Now transform that to AI. AI can rapidly synthesize information for review by the underwriter. Our AI agent does the underwriter’s work, reading every document plus relevant information outside the data room, and produces a report on whether we want to underwrite the risk—all in under 10 minutes, versus the better part of a week before. Our underwriter, who used to be quite tired after manually researching and producing a report, is now fresh and able to make a better underwriting decision. We’re also able to go after smaller deals and compete in areas of the market where others aren’t. Over the last 12 months, using AI, the business has grown more than 75% on a top-line basis. More interestingly from an insurer’s perspective, our loss pick—the expected cost of claims—has fallen by 5 percentage points. So if we’re writing $100 million in that area, better decision-making alone saves us $5 million, while the business has grown 75%. That’s transformational. And this hasn’t been about taking people away. In the last 12 months, we’ve added three people to a team of 10. They’re doing more profitable work and helping us build a stronger moat around what we’ve created in warranty and indemnity. It’s transformed how we understand and price risk and how we engage with brokers. And it’s given us a happier team: Underwriters spend less time on the grunt work of reading thousands of documents and more time negotiating and winning business. Q: Do you mind talking a little bit about Cortex by Markel, the problem statement, and how the team has tackled that? Wilson: Cortex by Markel is an AI-enabled business unit built by Markel with Bain that puts underwriters at the helm driving decisions. It assists in the underwriting of hard-to-place US casualty risks, things like hotels and restaurants in higher-crime areas, where legal-system cost inflation has made broad coverage unprofitable. Most insurers say no; we saw an estimated $10 billion opportunity. The problem was speed. Underwriters would get an inquiry from a broker, decide manually whether it was worth pursuing, then push it through legacy systems—a process that took one to two days. Market leaders were responding within four hours, and often the fastest quote wins the business. We had the distribution but not the speed. Rather than tweak our existing process, we challenged ourselves to reimagine it entirely, pairing Markel’s deep market knowledge with Bain’s strategic and technological expertise to build, in effect, a startup within a large corporate business. Within several months, we launched the first version into the marketplace. It has transformed the way we do business. Instead of an email coming in, a human being reading it, and using an old system to try to price the deal manually, the email now hits the inbox and the data is ingested automatically. The system quickly provides the underwriter with information on how to measure the inquiry against the risk we’ve articulated. While remaining fully in control of the process, the underwriter can now evaluate the inquiry quickly and comprehensively from every angle. Next, with the support of our approach, they can provide good, better, and best options to brokers who can then advise their clients. What used to take a day or two of manual work is now ready in 5 to 10 minutes, and back to the broker in under half an hour, with no change on their end. It’s a game changer: In the first four weeks alone, we quoted risks worth more than $1,000,000 a week in gross premium. It’s been a terrific reimagination of an area of business using a startup mentality, combining the best of Bain with the best of Markel to create something that could become the blueprint for our entire US wholesale and specialty business division. Q: What’s ahead for you and Markel as you think about your priorities related to AI for the next 12 to 18 months? Any business ambitions or aspirations it could help enable? Wilson: The companies that put AI at the heart of the business—using it at scale and redesigning processes end to end—will be the winners in five years. And it will be a compounding of the work that’s being done over the next 12 months. The starting gun has already gone off. Too many organizations are still tinkering with pieces of a process and calling that AI. The real opportunity is to build market-leading technology across each of our specialties at a reasonable cost. We need to grab that opportunity today and build an advantage that compounds over time. I think of it as having 100 startups under one roof, each improving over time as the AI mindset improves. Markel is almost 100 years old and writes about $10 billion in premium, but we’re still a relatively small player in the global market. The sky’s the limit for us. What’s exciting is that people across the organization, from those early in their careers to those with decades of experience, see AI as a massive part of our future.

Simon Wilson is putting AI at the center of how Markel Insurance competes by reimagining entire lines of business rather than optimizing them at the margins. As CEO, he is steering the company’s AI strategy across the organization, guided by a Founder’s Mentality that treats every process as one worth rebuilding from scratch rather than merely improving.

Kyle Kinsell, a partner in our Financial Services and Performance Improvement practices, sat down with Wilson to discuss the future of specialty insurance, how AI is reshaping the economics of underwriting, and why he believes the winners of this shift will be decided in the next 12 to 18 months.

Q: How do you envision AI transforming specialty insurance over the next several years?

Wilson: AI should be viewed in the context of everything else happening in the industry and the wider economy—geopolitical risk, marine war insurance, US casualty, legal-system inflation, changing US Treasury rates. All of it affects the specialty insurance we write around the world. But amid all that noise, I think there’s a danger of missing what is a deeply important technological shift. In my 25-year career, we haven’t seen anything like it. AI represents a fundamental shift in white-collar work, equivalent to what the Industrial Revolution meant for blue-collar activity.

From a strategic sense, we’ve worked with Bain a lot over the last two years, and one concept we’ve really picked up is the Founder’s Mentality. Founders are willing to do things differently from incumbent players. That Founder’s Mentality is at the core of our potential strategic advantage as a challenger in the specialty insurance market. AI is the defining tool that will allow us to think like a startup again and get back to that genuine Founder’s Mentality. It is very much at the heart of our business.

Q: Can you share how you’ve approached the AI journey at Markel so far?

Wilson: About three years ago, as large language models took off in our personal lives, we rolled out Microsoft Copilot to just over 5,000 Markel Insurance employees. In that early stage, there's always 5% to 10% of people who pick it up and run with it, but it doesn't have a material impact on the rest of the organization initially.

Where I started to see a genuine change within the business on the back of AI implementation was when people could define a specific and important business problem that they were trying to solve and when they showed that Founder's Mentality and owned the outcome. That’s been a great realization for me. If you start with a tool in search of a job, you’ve got a problem. If you start with a problem that could benefit from the use of these tools, you start to see things in a different light.

Q: What are the key business benefits you’re seeing from AI-enabled underwriting?

Wilson: One example is in the group that underwrites warranties and indemnities—coverage for buyers in M&A deals if something missed in due diligence causes a financial loss. It’s a popular product: Sellers like it because it gets deals done, buyers like it because it protects them, and boards like it because they’re covered if a mistake slips through.

The underwriter’s role is to insure this and price the risk. Traditionally, our underwriters—a group of buy-side lawyers—would spend about a week trawling a data room of thousands of financial and legal documents, giving maybe one in 10 a fair read. An underwriter could produce about four quotes a month, and we’d win roughly one in four. Given those economics, we tended to pursue only larger deals.

Now transform that to AI. AI can rapidly synthesize information for review by the underwriter. Our AI agent does the underwriter’s work, reading every document plus relevant information outside the data room, and produces a report on whether we want to underwrite the risk—all in under 10 minutes, versus the better part of a week before. Our underwriter, who used to be quite tired after manually researching and producing a report, is now fresh and able to make a better underwriting decision. We’re also able to go after smaller deals and compete in areas of the market where others aren’t.

Over the last 12 months, using AI, the business has grown more than 75% on a top-line basis. More interestingly from an insurer’s perspective, our loss pick—the expected cost of claims—has fallen by 5 percentage points. So if we’re writing $100 million in that area, better decision-making alone saves us $5 million, while the business has grown 75%. That’s transformational.

And this hasn’t been about taking people away. In the last 12 months, we’ve added three people to a team of 10. They’re doing more profitable work and helping us build a stronger moat around what we’ve created in warranty and indemnity.

It’s transformed how we understand and price risk and how we engage with brokers. And it’s given us a happier team: Underwriters spend less time on the grunt work of reading thousands of documents and more time negotiating and winning business.

Q: Do you mind talking a little bit about Cortex by Markel, the problem statement, and how the team has tackled that?

Wilson: Cortex by Markel is an AI-enabled business unit built by Markel with Bain that puts underwriters at the helm driving decisions. It assists in the underwriting of hard-to-place US casualty risks, things like hotels and restaurants in higher-crime areas, where legal-system cost inflation has made broad coverage unprofitable. Most insurers say no; we saw an estimated $10 billion opportunity.

The problem was speed. Underwriters would get an inquiry from a broker, decide manually whether it was worth pursuing, then push it through legacy systems—a process that took one to two days. Market leaders were responding within four hours, and often the fastest quote wins the business.

We had the distribution but not the speed. Rather than tweak our existing process, we challenged ourselves to reimagine it entirely, pairing Markel’s deep market knowledge with Bain’s strategic and technological expertise to build, in effect, a startup within a large corporate business. Within several months, we launched the first version into the marketplace. It has transformed the way we do business. Instead of an email coming in, a human being reading it, and using an old system to try to price the deal manually, the email now hits the inbox and the data is ingested automatically. The system quickly provides the underwriter with information on how to measure the inquiry against the risk we’ve articulated.

While remaining fully in control of the process, the underwriter can now evaluate the inquiry quickly and comprehensively from every angle. Next, with the support of our approach, they can provide good, better, and best options to brokers who can then advise their clients. What used to take a day or two of manual work is now ready in 5 to 10 minutes, and back to the broker in under half an hour, with no change on their end. It’s a game changer: In the first four weeks alone, we quoted risks worth more than $1,000,000 a week in gross premium.

It’s been a terrific reimagination of an area of business using a startup mentality, combining the best of Bain with the best of Markel to create something that could become the blueprint for our entire US wholesale and specialty business division.

Q: What’s ahead for you and Markel as you think about your priorities related to AI for the next 12 to 18 months? Any business ambitions or aspirations it could help enable?

Wilson: The companies that put AI at the heart of the business—using it at scale and redesigning processes end to end—will be the winners in five years. And it will be a compounding of the work that’s being done over the next 12 months. The starting gun has already gone off. Too many organizations are still tinkering with pieces of a process and calling that AI. The real opportunity is to build market-leading technology across each of our specialties at a reasonable cost. We need to grab that opportunity today and build an advantage that compounds over time. I think of it as having 100 startups under one roof, each improving over time as the AI mindset improves. Markel is almost 100 years old and writes about $10 billion in premium, but we’re still a relatively small player in the global market. The sky’s the limit for us. What’s exciting is that people across the organization, from those early in their careers to those with decades of experience, see AI as a massive part of our future.