Can generative AI truly transform the insurance industry, or is it just hype?

This was the fundamental question that underpinned Oxbow Partners’ interviews with executives and senior leadership at 22 of the largest reinsurers and specialty carriers in the world.

Perhaps surprisingly, the answer was almost unanimous. The vast majority of participants (19 out of 22) expected generative AI to transform the industry in the long term.

As a management consultancy specialised in insurance, we wanted to dig beneath the surface. What is this perception based on? How realistic is it? And what needs to happen to enable this transformation?

And this is where it became interesting. The ambition is not just hype. It’s not just a theory without basis. There are real use cases being explored across every part of the business – from underwriting to risk, from operations to legal, from claims to pricing and from portfolio management to finance – each with the the potential to be truly transformative.

It is true that some of the use cases being explored, are “merely” substitutive – their value is in replicating current activities using a new technology. It is worth noting that this type of use case is still valuable – however, some are far more transformative.

Think about the fundamental issue that all insurers are facing: the importance of synthesising the wealth of data that is theoretically accessible to insurers and reinsurers, to drive decision making.

Imagine a world where any and all data submitted, whether in number form, qualitative answers to required questions or general information embedded in a report, could all be synthesised into usable data and insights, with red flags (alongside rationale) generated to identify inconsistencies or other issues.

What if that data could then be augmented with public information about the company and the industry, at the click of a button?

What if any and all interactions with the client, including transcripts from meetings and email interactions from anyone at the reinsurer, could be used to validate, test and interpret the data already gathered?

That is just one possible application of GenAI. You might argue that this is a bit “out there”. But elements of these are already being explored by specialty players today. And with the exponential development of GenAI technology, how long will it be until the loss ratio benefit from improved decisions makes this a must-have, rather than the luxuries of leaders in the space?

This is still a while away for most carriers today – but let’s think about a claims handler. They are limited today in looking at, say, a couple of reports before having to make a decision on a claim. Imagine they were not so limited and could have summaries of dozens of reports and information from the public at their fingertips, to create a recommendation on a claim. How much better would the decisionmaking be? How much quicker would it be? A human remains central, but is given the ability to do far more.

This is not so far away for carriers. The summarising and generative capability of GenAI is easily accessible. And with most (62 percent) already developing or having a native Copilot or internal GPT in use, the GenAI experience is becoming very real on a day-to-day basis.

Examples not only of substituting activity but also of reimagining the possible show how GenAI goes beyond hype and into reality.

Is the insurance industry ready though? What’s holding it back? Why has nobody really developed a use case and scaled it across their business?

In our research, leadership per se was not generally a barrier (rated at 6.3 out of 10). There was genuine excitement from many and willingness to push forward, as long as the case could be proven. That is why many are still experimenting with proofs-of-concept as they navigate how to create the case for change.

However, there is a lot to be done and a number of challenges: the talent and GenAI skills necessary to take advantage of the possibilities are not native across insurers and reinsurers (lowest rated at 4.4 out of 10), and the culture that is needed to take full advantage of the potential takes a while to mature – and that’s with the push from the top.

In addition, there are technology and data foundations that need to be built – especially for those with legacy systems – that will hold back progress and create other strategic priorities in the near term.

And let’s not think that GenAI is without its faults. There are many challenges with the technology itself: from hallucinations to the quality of its outputs and its ability to scale.

But with the direction of travel identified, the potential of GenAI clear and the willingness to act already demonstrated, it is only a matter of time before GenAI becomes an essential strategic enabler for all insurers.

The question is: when?

Miqdaad Versi is partner at Oxbow Partners.