At Insurtech Insights, Lucy Pilko, CEO of AXA XL Americas,
This article is from a longer interview and edited for clarity.
Where does the industry stand with implementing AI?
We have been using AI. There's AI and then there's generative AI. We've been using AI in pricing models. As I said on stage, this is a data science industry, so it's not that AI is new to us. The process bit and the ability to do that better is at some level new. There's a lot of attention right now on what that can be. My attention and AXA's attention is really trying to focus on the places where it's not just process. It is actually going to be about insight, which means it's focusing on – yes, there's efficiency, but it's much more about what's the new value add in underwriting. How can we underwrite risks that today feel uninsurable? How can we use this information to better understand where we're a good fit from a balance sheet perspective? How can we use it to understand how clients can do more on a preventative basis?
How do you think about partnerships with tech providers?
If I go back 10 years and think about what happened with the initial rise of cyber insurance, commercial insurance companies had the claims data. They knew the economic impact. Technology players knew what was successful and what was not, what the threat environment looked like, and when and how defenses were successful or not. Getting those organizations to collaborate just never happened.
From my perspective, the world is getting too risky too quickly to not lean into those types of partnerships, because technology companies need and want their clients to adopt their technologies. Insurers can again grease the wheels of making that possible if the technology is understood, and the data is shared, and everyone takes a risk mitigation mindset. I think about how we build the communities, the forums and the opportunities to share data in a way that actually propels everyone forward, if we have access to that information.
How long will it take to accomplish these goals for AI?
We work in close collaboration with various LLM [large language model] providers. We work with the large cloud companies, to really understand a few things. One, what are they doing? Understanding their strategies, frankly, insuring them so that we can get even more comfortable with insuring companies that then use their technologies, is one of the initial steps here. But there should be opportunities to create forums where we can fuel adoption of technology, maybe through embedded insurance.
What can AI do for risk professionals? Will it enhance careers?
AI is making people bionic – the people that are willing to go on that journey to say, 'Okay, I now sit in the cockpit of information, and I can have more reach and control.' I think about our risk engineers a lot. They spend a lot of time on the road. They are physically traveling to visit plants. When you then give them that plus a set of tools that can look at the broader set of properties that a client owns, all of a sudden, if they embrace it, they have the potential to counsel and advise and manage risk with the client in a much more powerful way.
It's the ability to do more with information, because time is one of our most constrained assets. As we think about demographics, the world is producing fewer and fewer people every year. In that bigger context, we need each person to be more impactful from an economic perspective, and these types of tools help us do that. The mindset is how can we make each person more powerful than they are.
Will AI help people think more strategically?
On a higher plane, more strategically, more broadly. You can touch more things faster because you're not doing the non-thinking work that is associated with that. You're fed with the information to make good decisions. The challenge there is around skills and how you train people. In some ways, this creates tons of opportunity for people who already have a lot of expertise, because you give them more signals and they understand what to do.
What about the generation that's just graduating from college today? How do you help them understand what to do with those signals? It becomes a different challenge of, how do we bring those people into the workforce in a very successful way? On the flip side, they're much more comfortable with the technology. So they have skills that the more established colleagues don't have, but they also lack a certain amount of the experiential learning.
That's the challenge for us. How do we help people get up that learning curve faster? In call centers, when you put in generative AI technology that assists reps in answering people's questions, it often takes a year for people to get fully qualified. That's a very frustrating time for employees. You see a lot of attrition in call centers in the first months because they are uncomfortable giving advice. They know they're slow. The customers get frustrated. It's similar when we think about this more broadly, like, how do we get people up the learning curve? How do we make that a really enjoyable experience for humans?
