The Hidden Costs of Getting ‘AI Ready’

By Rosina Smith, Regional Director, UK at Sixfold

At Sixfold, we have worked with 50+ underwriting teams, implementing AI across carriers, MGAs, and reinsurers, over the last 3 years. One pattern keeps showing up: the teams that are succeeding aren’t the ones with perfect plans and data, they’re the ones who are learning while others are still planning.

Careful AI planning and strategy sounds smart in an industry like insurance. But after watching and being part of a lot of implementations, I’ve seen something that contradicts this: waiting isn’t free. 

They say, time is money, and while you’re planning, three things are compounding:

1. Every month you wait, competitors get smarter. AI learns from real submissions. The carriers who started six months ago have systems that understand their specific appetite, have been corrected on thousands of complex cases, and get better with every new submission. You can’t pay to skip that learning phase. Competitors are already ahead and pulling further away with every submission they process.
2. The expertise window is closing. Your senior underwriters are retiring. The ones who know which guidelines matter most, who can spot risks others miss, who understand decades of underwriting complexity. Carriers who started encoding that judgment and converting it to institutional knowledge months ago did it while those experts were still there to validate and refine. Every retirement that happens during your planning phase closes that window permanently.
3. Organisational readiness can’t be bought. The technology is implemented in weeks, but your team learning to work with it in the best way takes longer. Knowing when to trust the AI, how to challenge it, which judgments still need human review, that only comes from using it. Teams that started earlier are building that muscle while you’re still planning.

Meanwhile, brokers are already experiencing carriers who quote in hours with competitive pricing. Their expectations have already changed, they are not waiting for insurers’ implementation timelines. And the early-career underwriters you want to hire? They’re signing offers with carriers who already use AI daily, because working with it is the future.

So, what should insurers do? 

I’m not suggesting reckless speed. But there’s a middle path: start focused, learn fast, build from there.

The teams seeing results picked one bottleneck, often in the lines of business that previous technology couldn’t touch. The manual underwriting areas with messy data, complex risks, and 2+ hour research requirements per submission. That’s where the force multiplier is most visible. They started there, proved the impact with a clear before-and-after, then used that success to expand across other lines.

What they didn’t do: wait for perfect data, rebuild their entire tech stack, or try to solve every problem at once. The gap between planning and doing when it comes to AI has a cost that compounds every month. Perfect readiness doesn’t exist.