Insurers need to cross the chasm into AI or face being left behind

Insurance is an AI-laggard

The 2026 Sapiens Insurance AI Maturity Index shows a chasm between those that are AI-curious and the minority that have acted on the greatest technology inflexion point of our time.

How mature is the insurance sector when it comes to AI adoption and operations? To me, it’s clear that it has a long way to go. Insurance remains at an early stage in adoption for what is perhaps the most disruptive inflexion point in a lifetime. This may be understandable in an industry that from the outside is perceived to be risk-averse, but being a laggard brings its own dangers.

I believe, quite literally, that many of those who don’t make a firm commitment to AI-fuelled automation and business process re-engineering will be prey to acquirers, will see declining revenues and will experience a debilitating customer exodus. Insurers and the surrounding ecosystem need to wake up and smell the coffee. Insurers can’t use concerns over regulators and compliance as a reason for inertia. The time for gazing at PowerPoint decks, trying out proof-of-concept exercises and pilots needs to end.

In 1991, Geoffrey Moore’s book Crossing the Chasm described the positive impacts of being an early adopter when key changes, such as an enabling technology capable of driving a paradigm shift, occur. Nothing has changed since. Similarly, the first-mover advantage – think of how Amazon turned retail on its head by embracing the web while others waited – remains a powerful and demonstrably correct notion.

Dipping toes in the water versus diving in

But AI adoption itself is still at a stubbornly early stage and it’s important that customers root out AI reality and apply it to their own businesses rather than blindly buying into supplier claims. Agentic AI may be everywhere in the media as a force for autonomous process re-engineering but it is relatively scarce in the real world of insurance. Our research into AI maturity sees a delta between interest in AI and real-world AI deployment with 87% of senior insurance professionals globally saying that agentic AI is important but only 16% running it in production.

In part, this is down to a fear of change. There is also a need for capital investment. Certainly, some pilots have disappointed and there is confusion over supplier choice with hundreds of vendors promising the moon. But, whatever the reasons, I believe that those who hesitate are taking a bigger risk than those trusting this enormously powerful technology.

Those bald numbers showing a gap between belief in AI and action hide significant variations. Servicing and Fraud/SIU are relatively advanced, as befits a technology that excels at automating multi-step processes and identifying anomalies. But it remains very low in reinsurance.

However, we can say with confidence that these numbers will soon be redundant. Two-thirds (67%) of respondents expect fully autonomous AI in insurance within three years. We are going through a period of rapid change and, as ever with technology adoption, those who hesitate will have a lot to lose in terms of process efficiency, customer experience, faster time to market, reduced error count, employee effectiveness and ability to execute change.

Get the basics right

For those that ‘get’ AI and want to pursue opportunities aggressively, it is important to get the foundations right. That’s why, when we first designed the Insurance AI Maturity Index, we built it around four key pillars: cloud migration, master data strategy, AI governance and HR policies. Particularly on the point about governance, insurers should look at the example of the autonomous leaders and regions like the Nordics, who have been quick to embrace AI and agentic AI. They do not allow concerns about compliance to be a blocker for transformation. Over time, we should see rapid progression and modernisation in each of these areas, but today there are clear distinctions between those leading the pack in AI adoption (what we’re calling the autonomous leaders) and those lagging behind (the reactives).

Take the master data pillar. Many insurers have the advantage of sitting on decades of useful data but if that data is not clean, reliable, categorised, joined-up and open to analysis tools, then the results will be suboptimal and perhaps even dangerous. Logically, ensuring the right data strategy is in place before allowing AI to interrogate core systems will be key to success.

Complementary capabilities are also important in the other pillars of cloud adoption, AI governance and HR strategy. Hiring and retaining people who can provide ‘human in the loop’ oversight will be key. According to our 2026 Insurance AI Maturity Index, global data indicates autonomous leaders will invest more in upskilling staff compared to reactives (71% versus 35%) and hiring AI-knowledgeable talent (57% versus 20%).

For those that get it right, there are glittering prizes. More than three-quarters (77%) of respondents say they can identify new products and bring them to the market on average three months faster.

What are we waiting for?

Insurance may be struggling with what I have called the perfection paradox when it comes to AI, refusing to unleash it at scale until they are fully confident they can trust the technology. However, I strongly believe it is incumbent on vendors to show hard execution plans and demonstrable returns on a rapid basis to reassure customers about what they will get from AI. Automating tasks such as interrogating unstructured data is a good example of showing how AI can positively impact business processes.

I have seen a South African insurer which is considering a radical change that can reinvent the motor claims cycle. A customer can take a picture of a vehicle, while checks are made based on a historical knowledge graph using first-party and third-party data. The insurer establishes it’s a write off and payments are made in close to real time. The customer is happy because they have received a reasonable payment quickly while the insurer is happy because it has reduced the cost of delivering the service and can now count on a loyal customer for life.

Reflecting on the findings of the Index, experience suggests that some IT leaders ‘get’ AI but are wondering where to start with it. Others may have been disappointed by pilots. These are perfectly understandable responses to a breakthrough innovation, but the example of those that failed to cross the chasm during previous technology inflection points suggests insurers must work out if they want to be a first mover, a fast follower or a laggard. Given how quickly AI technology is maturing and changing the dynamics of the insurance industry, I would argue insurers can ill–afford to hesitate. Their very futures depend on those choices.

Paul Wheeler is CEO of Sapiens