EP 10 - AI as a Partner in Insurance: Innovation Anchored by Regulation
For a decade, the industry has been told to expect disruption. New entrants would bypass incumbents. Regulation would slow change. Technology would replace legacy systems and perhaps the system itself.
In this Episode:
Francois Jacquemin explains how AI can evolve insurance through integration rather than disruption. This episode highlights how technology, anchored by regulation, strengthens trust and closes coverage gaps.
That prediction missed something essential. Insurance is not disappearing. It is adapting.
A foundation that enables society.
Insurance is one of the most regulated and resilient parts of the global economy. It is not abstract. It enables everyday life. You can drive without financial ruin after an accident. You can borrow for a home. Employers can protect their people. Goods move across borders with confidence. This only works because there is a disciplined framework that balances freedom to operate with responsibility to pay. Regulation is not a slogan. It is the mechanism that turns promises into outcomes.
What AI really changes.
AI does not replace that foundation. It enhances it.
It improves risk selection and pricing by using more relevant signals.
It reduces waste by identifying overselling and coverage overlaps.
It closes gaps where clients are uninsured or underinsured.
It improves service speed at the moments that matter.
And it can translate complex models into a language a client can understand and repeat inside their own organisation.
Hyper-personalisation does not mean hyper-complexity. The goal is not a contract that nobody can remember. The goal is clarity. A product that fits a life stage, explained in terms a client can carry into a board meeting or a family conversation.
Integration, not invasion.
The future is a value chain where technology is embedded, not a battlefield where one side wins and the other disappears. AI providers, automation platforms, and analytics partners will integrate with insurers, reinsurers, TPAs, and brokers. That integration is already visible in health, claims, and distribution. The winners will be those who pair technical ambition with operational discipline.
The discipline of measures.
There are two real risks. Doing nothing. Doing everything. AI adoption should not be a laboratory without a hypothesis. It needs a vision, a roadmap, and proof. Clear KPIs across cost, service, retention, conversion, and claim outcomes. If investment does not measurably improve one of these, it is not yet ready for scale.
Why regulation matters even more.
Regulation is often treated as a constraint. In reality, it is the reason innovation can last. It protects the long game in a market that can be tempted by short horizons. It anchors solvency, consumer protection, and fairness. It also keeps trust intact when something goes wrong. Clients do not buy code. They buy confidence that the system will pay and that the outcome will be fair. AI must serve that confidence, not undermine it.
Human first, technology next.
Insurance remains a people business. Health claims involve fear and hope. Disability involves identity and dignity. Even property claims carry memory. AI can make answers faster and guidance clearer. It can keep a human available at the right moment, not as a last resort. The standard is simple. If the client feels more seen and better served, we are using AI correctly.
The way forward.
Insurance is here to stay. AI is here to help it become simpler, faster, and fairer. The path is partnership. Technology is integrated into the value chain. Innovation measured and governed. Regulation as the anchor. Trust at the centre.
Leaders who hold this balance will not only protect today. They will build the system that their clients will trust tomorrow.
Timecode:
00:00 Introduction to AI in Insurance
00:54 AI's Role in Health Insurance
01:40 Integrating AI in Insurance Companies
02:36 Client-Facing AI Applications
03:09 AI for Risk Assessment and Pricing
04:48 AI in Insurance Distribution and Consulting
06:54 Balancing AI Implementation
11:17 Future of AI in Insurance
13:45 Regulation in the Insurance Ecosystem
19:40 Challenges and Opportunities in Insurance Legislation
24:28 Conclusion and Final Thoughts
Francois Links:
Apple Podcast
Transcript:
So the topic we wanted to address now was, um, AI in insurance, or I should say AI and insurance. Uh, because AI is, uh, is there to stay. It's, uh, it's uh, it's a big hive for the moment. Uh, lots of money flowing in. Um, and it's been, of course, a lot of, of talks inside and outside the insurance world. About, uh, ai, about the use of ai, uh, about, uh, of course the danger, the pros and cons.
But, uh, one thing is important is that, uh, for insurance industry, AI is, uh, is a reality that they are facing as well as, uh, able to integrate, uh, to a certain level. And, uh, why do I say a reality that they are facing is because, um, around insurance, AI is very present. If you take, um. In, um, uh, health insurance for instance, uh, AI is extremely heavily used into developing a new treatment, new technology, new tools to, uh, help patients and, uh, insurers on the other side. Um, they are having insured people there being treated by, um, AI powered tools, so lack of understanding of that technology. Uh, would be wrong for an insurance company. It would, uh, it would do very much, uh, good for first client interaction, um, premium adjustment, and, and, and claim mastering to be able to understand how AI is being used by external parties.
Now this is one angle. The other angle is, uh, insurance companies need to understand also how they can integrate AI and, um. As we discussed yesterday, there was a, there's a lot of history, uh, in most of insurance companies, uh, who, uh, need to be able to find their right way to use ai. And there was, uh, of course digital revolution. There was big data for like 20 years or such a long time. And so it's been a lot of evolution, uh, towards, uh, better integration of technology and insurance industries. One of the heaviest investor in the in tech. Uh, not tech companies, well, I guess in tech companies as well, but also in, in tech services, um, and, and, and everything that goes around in the, in, in that ecosystem.
So, uh, AI is, is of course, uh, one element that, that, uh, can be of extreme use for, for insurance. I see several angle. The first one is, uh, is client facing. So how can we help? Um, clients, uh, experience to be improved thanks to, to ai. Um, but then in the background, uh, there's a lot of, uh, potential use as well. There is the efficiency. There is also, um, the, the best mastering of the risk environment where the insurance companies operating. By that I mean by having a lot of data. In the past there was a, there was use of data, but not an optimal use of data. Thanks to. The, um, the power that AI can bring and, and, and, uh, when, when ai, I mean AI together with all, uh, human experts and, uh, people who can steer AI to optimize the way data are assessed and, and, and the risk management is operated, um, then insurance company can use that to improve their own, um.
Risk assessment and therefore set the right price for it. But not only for that, um, but also to be able to explain to clients. That's when I mean that I, I mean, um, more corporate client who when they receive a price for a certain coverage, they are in need. Maybe, maybe, maybe need, but also in wish. To understand better why a price has been set in a certain way. And, um, the models of insurance, uh, are in arrangements are extremely complex. What, uh, AI can help is also there to translate the model into something which is easy to, and or easier to understand, uh, for a client, not from the point of view of the expert, but from the point of view of, uh, the person who knows.
A lot, but not maybe enough to understand the full modeling. And, uh, who needs to also explain internally as a, in, in, in that, in that corporate clients, uh, organization, why the price is certain in, in that way. Um, it's also, uh, not only for insurance companies or insurers, but also for the whole distribution suite.Um, there, there's a lot of, uh, actors in the insurance ecosystem. Who, um, on the one hand can make better use of, uh, AI for client optimization. But, uh, if I take all the consulting, um, companies and, and, and, and brokerage companies who help client to improve their own coverage, um, themselves, uh, also in need of, um, better finer, more advanced modeling, uh, that can be, um, used.
To allow the client to better understand and, and ask the right question for insurance companies. So, long story short, AI is a, a, a very, very good tool. It's, uh, but again, is a tool that need to be operated, uh, by, by humans, with humans in collaboration with humans, we could say. Um, and that will drive. Uh, strong improvement in the, uh, the insurance, uh, industry and ecosystem.So it goes, as I said, much beyond the insurancers is, uh, the whole client, um, environment, the whole distribution, consulting environment, the reinsurance environment, and, uh, the, the, the ecosystem around, uh, insurance, which is extremely, um, strong and, uh. Lots of new things are, are happening there. A lot of innovation, lots of great ideas, a lots of bright people are just, uh, finding new ways of, uh, servicing and, and, and helping, uh, already established actors to, to improve.
They're also competing. So there's of course, uh, beyond, uh, the improvement element. There's an element of danger there. And, um. I see several dangers. Uh, the, the first danger is not to do anything with ai. That's the first, uh, the second danger is to do too much with ai. Uh, so, uh, when AI is being implemented from a, from an insurance point of view, uh, it should be, uh, not as a, uh, laboratory to do, uh, experiment.And, and maybe something will come out of that. But, uh, there has to be an objective. Uh, we talk about vision when, uh, leading a company has to be a vision. Bring the people behind it. What is the same for implementing, uh, elements of AI within an insurance company or, or, or, uh, as part of companies or the ecosystem?
There has to be a vision. There has to be measures. There has to be the ability to demonstrate that investment is, um, done to bring a better outcome than, uh, before that investment, better outcome. Uh, it should be financial. It should be, uh, more client should be more client satisfaction, more client retention.
Um, so there's, there's a lot of, uh, KPIs that can be put on the, on the table there, but, uh, it has to be measured. It has to be also, uh, a positive outcome in the world. We are today, we are at the beginning of this revolution. We don't know yet where it's leading us, and there's still a lot of experimentation that are happening simply because the depth of experience of them. The whole, uh, people involved is not, uh, is not extremely long. However, it's improving all the time. And, and every step taken is, uh, is a, is a, is a, is a learning that, uh, drives these, um, these tools in the, in the right direction.
Let's take an example, for instance, uh, with ai and I will. Try to describe the world, uh, before AI where, um, insurance companies are organized in, in silos, uh, I, I mean here, insurance companies where there are a lot of, uh, different products that are offered to, to, to clients, um, which can be from the house, the, the car, um, health insurance, life insurance.Um, and, and, and I mentioned retail here because I would like to, uh, bring the example of a, of a person that I could be or whoever, uh, an individual client is where, um, as everything is organized in silos in inside the insurance company, creating bridges between, uh, the different type of product, uh, of course has been, uh, getting insurers very, very busy.
Uh, in, in the last, uh, the last decades to, to be able to, um, set an agent or, uh, the ability of a, of a, of a contact person in front of the client to discuss about and explain the different product that exists and, and, and that are available to the client or already, uh, in the client basis and in a, in a, the client is already applied or, and those, those contracts.But, uh, it doesn't maybe fit his own, uh, life situation at the time or his family situation at the time. Now, this, this is happening now, but it's still very siloed and very clunky in the way the processes are, are working behind the scene. And, uh, for the client, uh, what it, it it means is that there is a lot of hope.
So I had a, an agent in front of me a few, few weeks ago. And he knew my, my contract, uh, he didn't know everything about my contracts, but he knew a little bit and, uh, there was a lot of expectation that was created. Now, what AI will allow is that these expectations are met because of the, of the, of the silo and the loneliness of the, the, the clunkiness and of, of, of those tools that exist, uh, simply because they were built in a, in a, in a separate way.There is always a delay between. Um, the moment the expectation is created and the moment where things are happening, and AI will be able to improve and reduce that time from weeks to days or even, um, minutes or a second to allow a, a much quicker way of interaction. Um, that's, that's an example that I, I could, I could take.
And that leads me to what, what this industry will look like in the, in the future. Um, there will be much more personalization of the insurance and, um, offering, uh, which be, be, be better adapted to, to, to the situation of a client. Of course, it requires exchange of information. But, uh, the ability of an insurer to hyper ize is going to, to change the game.And, and then, uh, it doesn't mean that the situation today doesn't, will not evolve. That's the whole point, uh, is that there is a life cycle understanding and that the insurer demonstrate the ability to follow that lifecycle. There's a lot of misunderstanding about hyper-personalization.
Hyper-personalization doesn't mean hyper complexity. It doesn't mean that every client has a contract which is extremely specific to his or her situation, and therefore create an extreme complexity in a back, uh, of the, or in the back office or in the, in the, in the administration. What it means is that there is a clear understanding of.The situation of the client and that the offering is packaged in a way which is clear and understandable by the client. So not necessarily having so many options that the clients even can't remember, uh, what the options are of a contract. Several, um, minutes or, uh, hours or days afterwards. Uh, but something which is transparent, which is, um, trustworthy.
Uh, which is adapted so that the client feels that there is an understanding of his or a situation and the ability to respond to those needs in a, in a very quick way. AI will help bridging all those silos and, uh, making, ironing out, uh, differences in the way the, the, the service is being delivered to the client.And, uh, also will allow for a 24 7 availability of. Element of answers and q and As and maybe avatars and chatbots, but evolve chatbot, not, uh oh. Thank you for your question. We'll refer you to, uh, to an agent in, in the next 48 hours, but, uh, a, a, a tool that allows a client to receive already quite a lot of information whenever he has some questions.
Regulation. Big topic for the insurance ecosystem. Regulation is the basis of the insurance ecosystem. It's what's creating, uh, the insurance contract, which is the bond between a client and, and insurer, and, uh, establishes, uh, a, a more tangible truth, uh, element or a, a, a, a catalyst for truth between the insurer and, uh, and, and the client, or the client and the, and the insurer.
What does that mean is that the insurer, uh, receives money from a client who pays without knowing if he or she will ever be in need of that, and on the promise that the insurer will pay, uh, in case of, uh, of damage car. It could be fire, it could be an iPhone that's falling on the floor, um, and breaks down, or, you know, smashes or whatever is, is, is happening, uh, that.Drives, uh, a claim to be filed with the insurer and the hope and the trust that, uh, the insurer will pay the due amount. There's always, uh, issues there. Uh, and, uh, there's always more expectation from the client and, and, uh, the insurer tries to do, uh, usually the, the, the right job, uh, paying the right amount based on, uh, the, uh, the contractual conditions.And those conditions. Of course, they're regulated by, uh, a, a set of legislation. Insurance is often seen as a very traditional, little bit old fashioned in the way that, uh, from the client point of view, this, uh, regulation, um, from this point of view are not changing. The contract are the same, uh, for the last, uh, decades.Uh, the, the general conditions are still called general conditions. Uh, the application form is still called the application form, the civil medical questionnaire. Or you still need to give the, the power of the car and, uh, the, the, the, the size of the, and the color of, uh, the, the, the car, et cetera. So there, there's a lot of things that aren't changed because the materiality of the, the situation doesn't change.
And therefore, from the client point of view, there is no change from the insurer point of view. However, uh, there has been so many changes for the last, um, uh, let's say 30, 35 years. In the, in the insurance industry, um, it's, uh, it's first a facilitator. Without that piece of legislation, I don't think the insurance company would be able to, to survive because, um, it sets the rules for a long-term view and it balances, uh, the expectation from shareholders and, and, and, and everybody to achieve extremely quick results.Uh, we see a lot of impatient people in insurance. Um, but without this, uh, long-term need, establishing the legislation and, and, and the, the, the sets of controls, then, uh, there could be a lot of, uh, dangers, whether the financial or too risky or losing clientele, losing even, uh, the trust. So it's, it's a necessary catalyst to have a, a, a regulation.
It's also a necessary catalyst. Uh, for the long term sustainability of the, of the business. And, um, there is always the danger and it's something that we see, um, in Europe, but not only in Europe. And, you know, we, we, we, we tend to, um, be very critical about sales, but uh, insurance regulations are, uh, are very strict in, in many other countries, uh, you know, US or China, for instance, have also very strict.And, uh, and complex regulations in any way. Um, the, in, in Europe, uh, it is something that, uh, create for the public also, uh, a very strong, um, trust in, in, in the sys the system. Let's say that if you, in Spain, a complete change of the distillation to copy the French or German law, or Slovenia law, or whichever law.
There will be a big question mark. The insurance legislation cannot be dissociated from how the full set of society is working. Um, I don't consider insurance as a pillar of society, but basically it's a basis that allows the pillars of the industry and and, and the economic world thrive in the, in the, in a, in a society.And, and therefore it's extremely closely interrelated with all those, uh, all those elements. Uh, you, we, we, we speak about social security or even car regulation, um, or so, so many element of the, of the legal system that, uh, it, you can't change a, an insurance piece of legislation fundamentally without impacting the whole society.
So that's why. There are a lot of, uh, difference between the, um, legislation of insurance in, in Europe, what, uh, is happening. Also, of course there are some, some ZI lots, uh, in, in the insurance industry, uh, legislation and, and, and, uh, control system that, um, tend to overregulate, but that's due simply to. Uh, to society as well. In society. You have a part of society that really wants a lot of regulation and a lot of state intervention and, and part of society that really wants to minimize that. Um, but because insurance is such a pillar of society, then the, the people part is always, uh, very well, uh, considered. And therefore, legislation has been, uh, has been growing significantly.
Uh, for, for, for, for a while now. And, um, it could be also challenging, uh, to implement that because as, as insurer, you would like to do business. You want to develop your, your, your services. You want to, you want to create something which is good for your client. You want to be there for them. You want to add the service element. Insurance is not only about being insurance now, it's also about all the service element that you provide next to the insurance. Uh, coverage and, and contract to be able to help the, the, the clientele. And, uh, often you are not allowed to offer the full extent of the service because of, uh, the, the element of legislation.
GDPR, for instance is, is has been established for, uh, at European level. And then of course each country have implemented in. In a more or less forceful way, depending of the, of the culture and, and, and, and government of the country. But it's something which can be extremely limited. Uh, just going to take an example, uh, an executive of a, of a very, very big, uh, English company, a client of ours, um, eight, nine years ago was, uh, had a a, a health problem and was, was disabled at home.
It was in the uk. So, uh, culturally and illegal in the uk there's a lot of services that exist to bring people back to work. So high level executive, extremely, you know, extremely well paid and, but also the way the pay was due to the extremely good competence that this person was bringing to, to, to his employer or employer.And, uh, after six months, thanks to the whole ecosystem, being able to being, being able to work. Using data, allowing insurance and servicing companies to work together. Uh, she was back at work with two key elements, but the most important was that she was so happy because she could have a life back. She was not somebody who wanted to be disabled and, and, and do nothing at home.
She was somebody who wanted to, to contribute and do something. Of course, uh, her perspective on life has changed afterwards, but it doesn't matter what hap, what matters is that the happiness of that person. From being a just miserable and feeling useless to somebody who was able to contribute again and, uh, have a social life and, and back and, and, and, and working life back there.So that, that was extremely positive. Of course, financially, uh, it was very good for the employer, for the insurer and, and for a social system. But in all, in all, you have a bad situation with a person who is not happy and you create something, uh, positive. And, um, sometimes legislation does not allow that to happen.
And, uh, that's something which also is the dark side, let's say, of the, uh, the, the, the consumer protection is, uh, is the overdoing it, the over patronizing of, of society. But hey, you know, that, that's my opinion. Um, because it's, uh, it's, I see this more as a supporting people and, uh, and, and, and of course there are others, uh, who will say, yeah, you, your.Personal opinion. Francois is a dark side as well, because sometimes insurer will tend to overuse that to be able to pay less claims. So yes, of course there has to be a balance, there has to be a, an understanding of, uh, everybody's situation. But I think legislation should remain as simple as possible, should still allow for helping people and not overprotecting, uh, on, on the on, on.
Blindly a situation which is neither good for the client, nor nor society. And, and, and I can't remember where I heard that, but when, when a a, a legal system is so complex that the whole population does not even under understand the, uh, the, the, the legal system anymore. And, and needs, you know, experts and, and field experts, you know, legal experts.They have all their field of expertise. To be able to understand then, then it's a moment where we need to think about what legislation is the right one, the right level, um, and, and be able to create the, the right balance between, between, uh, making sure that the industry is, uh, sustainability the long term that the client element remains the core of the, uh, the industry.
But that, uh, helping client is also something which, which remains allowed, uh, as part of the, uh, of the legal system. So, big question. No answer. I think it balances, uh, depending on, on legislators and, and, and, and orientation there and the political, um, situation. But, uh, it's something, it's a question that I will, I will never see solved and I, I don't believe anyone will, will see this one solved completely in the future.There's going to be always a, a challenge there.