Groundhog Day: why AI isn't transforming insurance (yet)

If you're a tech leader in insurance, you probably feel like you've lived this before. You've made serious headway with new technology. You've got a few use cases and moved a number of pilots into production. Boxes are ticked, and things seem to be humming along nicely. Yet when you look at the metrics that matter to you and the business, such as expense ratio, broker engagement, or quote lead time, there's a weird sense that everything is kind of the same…

Our new research shows AI is following a similar pattern. (opens in new tab)Every insurer we surveyed has put AI into production, yet only 23% describe it as transforming their business. So what gives?

Sometimes excitement gets the better of us. Most of us suffer from FOMO. It's easy to get caught up in a new idea and pin our hopes on something novel, because the everyday stuff is grunt work. New is fresh and alluring, and we're all susceptible to it; seasoned leaders and boardrooms included.

I remember when Duck Creek launched, and everyone decided it was the PAS they needed. We spent a couple of years implementing the technology, descoped things along the way, told the operations teams it was coming, added manual workarounds they still needed, and then launch day came.  
 
We've migrated to the new PAS! Great, are things better? Um…  

Sound familiar?

I expected things to be different with AI. I hoped we had learned our lessons. But our research found the top barrier to scaling AI was leadership and cultural resistance (39%), ahead of the usual technology hang-ups such as legacy integration, data quality, and talent shortages. As one respondent, a CIO, put it:

"We've invested heavily in the platform, but many business teams are still hesitant to change their practices."

That refrain could apply to any technology in the last few decades. We need to ask ourselves why this keeps happening.

Away from the hype, there's nothing new about AI. It's automation: another technology we already know how to land. And we know what makes it land: customer needs and user adoption. Ignore those and the project is dead on arrival. We're applying AI to processes designed for a different time and technology, so we chase efficiency instead of rethinking the work. If you want real change, you need to invest across the board: not just in the people who deliver it, but in bringing everyone else with you.

When I started in insurance, tech was a black hole to the business. Business leaders who didn't trust IT bought their own tools or built their own spreadsheets. The thinking went: you never gave me any technology to do this part of my job, and I had no budget, so I built it in the only thing I had.

That legacy is still with us. Our report found that 84% of insurers still run a fifth of their analytics on spreadsheets, even though 71% rate their data governance as advanced or leading.

Tech and business leaders have since made great strides, and tech strategy now traces back to business strategy. And yet AI is picking at old wounds.

We all know the fears about AI replacing jobs. We've seen this before. I’ve seen well-intended efficiency programmes deliver six-figure savings but result in unintended headcount cuts. The result is teams stop letting you in.

If every efficiency gain turns into a headcount cut, people will avoid the tool and quietly carry on as before. They need to see the time saved going into better work, not out of the door.

AI can reach into that space between programmes and BAU. And it doesn't have to be about reducing headcount. Of course, every CFO wants a better expense ratio. But you can get there by growing, not cutting, if you free people from drudgery so they can do value-adding work.

I know of one company doing interesting things. In a pilot, its cyber underwriters review AI quote recommendations instead of keying data into multiple systems, and the AI turns a submission into a recommendation in seconds. The underwriter does the interesting part: deciding whether to release the quote. That's the real power of AI: it moves business metrics and augments people, winning the hearts and minds of both your workforce and the C-suite.

If we already know the lessons, why do we keep repeating them? And what do the firms who get it right, the 23% club, do differently?

Check out my article, AI fluency and adoption: how insurers can make their AI transformation stick (opens in new tab), where I explore how we can break out of this cycle.