90% Want AI. 22% Have It.
90% Want AI. 22% Have It. Here's Why They Want It, But Don’t Have It
Nine in ten insurance execs call AI a top strategic priority.
Only two in ten feel remotely prepared to actually do anything about it.
The Gap Between Aspiration and Execution
I've been working with InsurTechs long enough to recognise the pattern. All Insurers know transformation is coming. They read the same McKinsey reports. Attend the same conferences. Nod along to the same keynote speakers banging on about disruption.
Then they go back to the office and... nothing changes.
The numbers tell the story better than I ever could. 90% of insurance executives identify AI as a top strategic initiative for 2025. Sounds exciting until you realise only 22% actually have AI solutions in production. The rest are still stuck in what the industry politely calls "the exploratory phase."
Which is corporate language for "we've hired some consultants and nobody's quite sure what happens next."
Here's the bit that makes this fascinating: it's not ignorance. C-suites genuinely understand the stakes. They've seen the projections showing AI-powered underwriting can cut operational costs by 30% whilst boosting customer acquisition by 20%. They know claims automation can reduce processing costs by 30% with near-zero error rates.
They know all of this. So what is stopping them to actually do it?
What 38% of them are actually afraid of
I was doing some research and found some interesting data.
38% cite "fear the technology will fail to deliver" as their top implementation barrier. Which sounds like a technical concern but isn't. Not really.
I've personally witnessed this in conversations. What they're actually saying is: "I don't trust that this will work in our environment, with our systems, with our people, and I definitely don't trust it enough to bet my career on it."
Because it ultimately always boils down to one thing. Someone—usually someone in middle management who's been at the company for 15 years and knows where all the bodies are buried—has to champion this thing. Has to stand up in meetings and say "yes, we should absolutely rip out this system that's been running perfectly well since 2008 and replace it with something from a startup in Shoreditch."
That person is taking on enormous personal risk. If it works, maybe they get a pat on the back. If it fails? They've just become the “yeah but, remember what happened to Steve” story that gets mentioned in every future innovation discussion.
The numbers add proof to this. When you look at where resistance actually comes from, it's not the front-line employees. They're only 6.45% of the problem. It's department heads (16.77%) and middle managers (11.61%). The people whose jobs and influence are directly tied to the systems everyone wants to replace.
The other barriers nobody admits to
Beyond the fear of failure, there are the practical nightmares.
Legacy system integration isn't a technical problem. It's a political one. 70% of IT budgets are consumed just keeping the old systems running. Which means there's no money for new things, and the people whose expertise is in maintaining those old systems have every incentive to explain—in great detail, with charts—why integration with your shiny new AI platform is basically impossible. I know many of these types… It always starts with “Yeah, it all sounds great but remember…”
Then there's the skills gap. The workforce that actually knows how insurance works is aging out. The new hires who understand technology don't understand insurance. Which is a complex bridge to join up.
Add in regulatory complexity, compliance concerns, and the general exhaustion that comes from years of failed transformation projects, and you've got an industry that's simultaneously desperate for change and completely unable to execute it.
What actually works (and why most won't do it)
The Insurers who've actually managed to bridge this gap—the 22% with AI in production—didn't do it by being braver or smarter.
They did it by making the change feel manageable.
For InsurTechs, thinking through the below when putting together a pilot proposal will significantly ease friction:
Start small. Ridiculously small. Not a "strategic AI transformation initiative" that touches everything. A 90-day pilot that solves one specific, painful problem. Something like "reduce claims processing time for routine motor claims from 14 days to 3 days."
That's it. One problem. Clear metrics. Fixed timeline. Success or failure becomes obvious, not something you can spin in a deck.
The pilot needs a pre-agreed path to production. None of this "we'll evaluate the results and then decide" rubbish that leads to pilot death. The agreement from day one is: if we hit these metrics, we roll this out to the whole department. End of discussion.
You also need to address the people problem upfront. The person championing this internally needs ammunition. Not just ROI projections (though those help), but proof that this won't be a career-ending disaster. Case studies from similar carriers. Security documentation. Compliance sign-off. Integration plans that acknowledge, rather than dismiss, the complexity of their legacy environment.
Make it so the champion looks brilliant, not reckless.
What this means if you're selling to them
Here's the truth: most insurers reading this will nod along, agree completely, and then do precisely nothing.
Because knowing what to do and actually doing it are entirely different things.
The gap between 90% aspiration and 22% execution isn't a knowledge problem. It's a courage problem. A politics problem. A "nobody wants to be the person who broke the thing that was working" problem.
And if you're an InsurTech trying to sell into this environment, understanding that gap is worth more than any product demo.
The insurers who actually buy—the 22% with AI in production—aren't smarter than the other 78%. They're not braver. They're not more innovative.
They've just found internal champions desperate enough for a win that they'll push through the bureaucracy. And more importantly, they've found InsurTech partners who understand that selling to insurance isn't about having the best technology.
It's about making the decision feel safe enough that someone will stake their career on it.
That's the game. You're not selling AI or automation or efficiency. You're selling career protection to a middle manager who knows that if your platform fails, they're the one explaining it to the board.
The InsurTechs who win aren't the ones with the cleverest algorithms. They're the ones who show up with ROI projections, integration plans, security documentation, and case studies from similar carriers—all packaged in a way that makes their internal champion look brilliant instead of reckless.
Most InsurTechs won't do this. They'll keep building features, perfecting demos, and wondering why deals stall in "exploratory phase."
The ones who get it will spend less time on product slides and more time understanding the political nightmare their champion is navigating. They'll realise that every "let's schedule a follow-up" is code for "I need more ammunition to convince the six people who can kill this deal."
If your deals keep dying in pilot purgatory, it's not your technology. It's that you're solving the wrong problem. The problem isn't "how do we automate claims." The problem is "how do I champion this without ending my career."
