Eliot Partnership’s Senior Advisor Spotlight is a series of conversations with the senior advisors who have spent their careers inside the markets we recruit for. After opening the series with Andreas Thiele in Asia Pacific, we turn our attention to digital transformation and AI in insurance.
Our second conversation is with George Kesselman, Special Advisor, Technology and AI, based in Singapore. George built his early career at some of the industry’s biggest names, in senior product and strategy roles at Prudential and AXA Life, and at AIG as Regional Vice President of Claims Operations and Chief Operating Officer, Indonesia. He then left corporate life to found the InsurTech Asia Association and his own start-up, which was later acquired, before leading the commercial office of ZA Tech, the technology business of ZhongAn Online. Today he is Managing Director of Willow Ventures, a guest professor at National University of Singapore Business School and author of The Power of Play. That mix of incumbent, founder and digital insurer experience gives him a broad view of how technology is changing the industry. He joined Eliot Partnership in 2022, where he advises Asia Pacific and the wider global business on technology, digital and AI.
In this interview, George reflects on why the claim is the real moment of truth, what ZhongAn understood that traditional insurers didn’t, and the themes driving deals in Asian insurance. He explains why the insurers getting real value from AI start with the problem rather than the technology, what boards most often get wrong when hiring a chief digital or AI officer, and why today’s insurance leaders need a founder mindset.
You held senior roles at Prudential, AXA and AIG, including COO Indonesia. What did running operations and claims at scale teach you that you still draw on today?
GK: The claim is the moment the customer is actually buying. It’s the moment of truth. Everything before that is a promise with a legal document attached. So it made me think about how you align the whole proposition, from how it’s sold and positioned to how it’s serviced, all the way through to the claim. The best way to sell is to make those moments of truth really delightful. That generates positive word of mouth, and it accumulates and multiplies. Get it wrong at the claim and the cycle runs the other way.
It also changed how I see operations. People tend to point fingers at operations as the source of problems, but operations is usually just where problems from every other area surface, from underwriting to sales. So it taught me to look at problems and solutions end to end, rather than fixing the tip of the iceberg that shows up in operations. I see the same thing in today’s conversations about AI. It isn’t about applying AI to one operations problem. It’s about linking up all those parts so everything flows.
What surprised you most about the move from incumbent to founder? What did you get wrong at first?
GK: The sheer speed of change, and how much I underestimated compliance and regulation. In a corporate role, compliance feels like friction. It takes time, but somebody else is handling it. As a founder, everything comes to you. You can build technology quickly and move fast on a lot of things, but compliance and regulation bind that speed. You want to experiment, and that cycle of experimentation slows right down. In a corporate you tend to think compliance is overdone. Then you realise some of those things really are hard constraints.
The other surprise was focus. In a corporate you’re given your job description and a machine that keeps spinning and producing the same results. In some ways that’s a lot simpler, because the boundaries are clearly defined. In a startup, your attention can go to a million different things, but your time and resources are limited. You don’t have a team of a hundred people, so you have to be very disciplined. Realistically you can do one or two things well, and everything else becomes a little less good. Time is the death of a lot of startups, so you have to keep reminding yourself that you have a fixed runway and ask how you get the most value out of it.
You led the commercial side of ZA Tech inside one of the most successful digital insurance models in the world. What did ZhongAn understand that most traditional insurers didn’t?
GK: They went where the customer already was. Rather than selling through agents and traditional channels, they went into what they call ecosystems: the e-commerce and payments platforms where people were already transacting online. Because the insurance was digital, the marginal cost of issuing a new policy was very small. So they could embed very low-cost cover at the point of transaction and push it at scale. They understood that the unit economics were completely different from selling through agents and brokers, where products need to be bigger and more complex to pay for those channels.
The other thing was that technology wasn’t a separate department sitting off to the side. It was part of the organisation’s DNA from the ground up. At one point more than half the staff were engineers, the reverse of most insurers. Other insurers could copy the products, but they couldn’t copy the speed of iteration or the DNA, and without that they couldn’t replicate the economics. That’s why the technology was later spun out into ZA Tech, whose international business is now Peak3, to sell it to other insurers.
In 2018, you predicted a Spotify moment for insurance, with products unbundled and embedded into everyday digital services. How close have we come?
GK: About halfway. Embedded insurance has really taken off since then. Short-tail products like travel, device protection and e-commerce cover found their niche and grew very successfully. But bigger, more complex products like life and health are still largely sold through advice, with a bank, agent, broker or adviser in between, and that constrains how they come to market.
So there’s a natural limit to how far embedded can go just by making products smaller and simpler. The next big shift, which is happening now, isn’t about embedding. It’s about AI making good advice cheap enough to scale more complex products to much wider segments of the population.
Through Willow Ventures, you work with insurers, banks, brokers and private equity investors. What themes are driving deals in Asian insurance right now?
GK: In Asia the theme is still largely distribution and growth, more so than in mature regions where the focus is on optimising profit. Here it’s about the next phase: increasing penetration, widening coverage and closing the protection gap. Distribution has historically meant bancassurance, agents or brokers, with agents the dominant force, but over the last five to ten years agency has started to saturate. So a lot of deal activity is about partnerships, particularly with banks, giving insurers access to a much larger distribution footprint.
Private equity is following a similar theme. The last wave was PE buying insurers to unlock value, but the lesson of that cycle was that buying insurers is hard and very capital intensive. So there’s been a big shift to balance sheet light models built around distribution. There’s a wave of investment into brokers, and on the life side into independent financial advisers, through roll-up plays where PE combines multiple firms into a regional group with scale. Managing general agents, which are very active in the US and Australia, are now coming to the region too. They underwrite the risk and pass it to insurers, and investors are looking to combine them to build a larger footprint.
What separates the insurers getting real value from AI from those running pilots that never scale?
GK: I don’t think anybody has massively cracked it yet. Most are still deploying early pilots and looking for ways to realise the value. Some of the most interesting examples I’ve seen are coming out of China, where Ping An and ZhongAn are front runners, embedding AI into the customer experience and into more customised agent propositions.
What separates the ones seeing value is that AI isn’t an experiment on the side. It’s part of the core scorecard for the executives and the business. The first phase of AI was a lot of “why don’t we try this”. Too often it’s still a solution looking for a problem, and that rarely creates value. The insurers getting it right start with a big problem they’ve never been able to solve. They don’t get bogged down in fitting AI into existing processes. They rethink the problem from the ground up: if we had no constraints, what’s the most efficient way to do this, and is that AI?
Something I’ve seen work particularly well with Chinese insurers is running three parallel streams, human only, hybrid and AI only, and letting the competition decide. There’s some waste in running duplicates, but you quickly see which approach is really winning. If the technology isn’t ready, they trim that stream and try a different route.
As AI takes on more of the work, what changes in what a CEO or senior leadership team needs to be good at?
GK: Capital allocation becomes critical: which tech bets to back and which to kill early. Historically the insurance CEO role was fairly predictable. Grow distribution, control operating costs, manage the investments. That’s going to change drastically as businesses become much more technology driven. And not everything should be solved with technology. Right now, trying to do that would be far too expensive and inefficient. So leaders need to calibrate the balance as the technology matures and new solutions come in.
The other big shift is talent. The culture has to change to make room for this kind of innovation, so leaders need to judge who will adapt to that shift and where they need to bring in new capability. They also need to attract the people who will be critical going forward. I’m hearing very active discussions from insurance CEOs who want a core technology team but are competing with banks and technology companies for the same people. Insurance is often seen as slow and a bit boring, and a really good AI engineer has everybody lining up to hire them. So how do you get them excited about being part of the change?
At the recent McKinsey Asia AI Summit, they called it a founder mindset, and I think that’s right. During this period of change and uncertainty, CEOs need to roll up their sleeves, lead from the front and work closely with the business to adjust processes and decide which technology to prioritise.
Boards are being asked to oversee AI strategy and risk. Do most insurance boards have the expertise they need?
GK: Candidly, most aren’t there yet. Boards generally have very strong capability on risk, but on technology there’s still a lot of early education happening. I’ve spent time with a number of boards, and everybody’s asking slightly different questions as they try to understand it. Many are being reactive. They see news about the latest AI development, then news about new cyber risks, and respond to each rather than thinking about it systematically.
That’s where there’s probably a need for more technology expertise on boards. Not to get too deep or too prescriptive about how the technology should work, but to put the right guardrails and governance in place so the company’s AI investments, roadmaps and strategy are heading in the right direction.
When a business is acquired or brings in private equity, what usually happens to the leadership team, and what should boards plan for?
GK: The main changes tend to be in the CFO and COO seats. Private equity wants to accelerate value creation, and that comes down to financial and operational discipline. The tempo under PE ownership is very different. With a typical five to seven year window to get the business to a much higher valuation, finance and operations have to work much harder. PE firms also often bring in their own technology teams and set-ups to help the business accelerate. Beyond those key roles, I don’t think there’s a huge amount of change.
For boards, the priority is succession and retention of key talent. As the pace picks up under PE ownership, some people will be worried by it, while others will be energised. So you naturally see some turnover during the ownership period, and having a plan in place ahead of time keeps the business in the best shape.
Your role at Eliot Partnership focuses on the firm’s digital and InsurTech offering. What does that look like in practice, and what do clients most often ask you for?
GK: The most common question is: what does good look like in this role? Clients often frame it as a choice between hiring from inside the industry or outside it, but that’s sometimes the wrong question. It’s really about defining what good looks like, what you’re actually solving for, and how that matches the maturity of the company.
Put someone brilliant from a tech company into a business that’s a long way behind, and you might only keep them for a short time, because they get frustrated and leave. So it’s about matching the right people to the right businesses, and finding people who know how to shape the technology but are also credible within the business, so they’re effective once they’re in the role.
In practice, that usually means refining the client’s brief. We make sure it captures the right elements of good, and that it’s attractive to the right candidates, because they’ll be looking for certain things themselves. The aim is a natural match between the right talent, the right company and the right framing for the role.
When an insurer hires a chief digital or chief AI officer, what do boards most often get wrong in the brief?
GK: The most common mistake is a big title with no P&L and very little say over process or how things actually get deployed. You need someone who understands technology and can lead change, but most importantly, can lead the transformation. As I said earlier, technology is only part of the solution. For the business to see value from that investment, its people and culture have to adopt the technology and thrive with it. Without a transformation mandate, even a very good technologist can land flat, and what they build becomes a bit of a white elephant.
The other thing is timing. This isn’t a short bet with instant results. Expecting results within 12 months probably isn’t realistic. Eighteen to 24 months is a more realistic time frame to start seeing them. As a new leader comes in, builds credibility and drives the transformation, results will lag a little behind. So in the meantime, boards should watch how the transformation is progressing rather than expecting results that move the needle before the 24-month mark.
What advice would you give a senior insurance executive who wants to stay relevant as AI reshapes the industry?
GK: Adopt a founder mindset. That means using the tools yourself, every week. It’s an extra time commitment to learn them and work them into your routine, but it gives you a real appreciation of what’s possible and where the limitations are. Rather than reading the news about what’s going to transform the industry, get hands on.
Second, get one live AI initiative under your mandate, something that touches your actual work rather than personal use, so you see the whole process end to end: starting it, getting it into production, and seeing the results and the limits.
And finally, spend time outside insurance. Get ideas from other industries and see what else is happening. Insurance can be a bit of a bubble, and if you stay inside it too long your worldview narrows. Getting outside really opens your eyes.
