After identifying the resistance, here’s what insurance leaders need to do right now—from executive strategy and data fundamentals to getting employee buy-in

Billions of dollars evaporated from insurance firm valuations after Spain-based digital insurer Tuio announced a deal in early February with OpenAI. With this arrangement in place, consumers could now receive a personalized home insurance quote and purchase a policy entirely within a single ChatGPT conversation.

The market reaction was swift and brutal. Insurance stocks fell by as much as 25%, Barrons reported. For insurance and other industries, the stock market’s dramatic reaction to ChatGPT’s insurance news was one of several shots across the bow for executives managing the latest round of digital transformation. 

Software stocks plummeted when ChatGPT rival Claude released new capabilities. Transportation and logistics faced similar sell-offs. The pattern has become clear: one AI application emerges in a sector, and Wall Street panics.

This is absolutely part of the hype cycle. Yes, those stocks will recover if they haven’t by the time you read this. But the real takeaway is that this is just the beginning of AI, and companies who dismiss it as temporary volatility are missing the point entirely.

It’s similar to the early internet. Remember the dot-com bust? Real business models eventually emerged, proving a fundamental truth: the internet was here to stay. It revolutionized every single industry; companies that didn’t take it seriously didn’t succeed.

Change is slow until it’s not. That’s one of my favorite maxims. And right now, we’re at an inflection point.

In my previous TIL article, I explored why resistance to AI persists in insurance underwriting: the expertise threat, data quality deflection, and sacred cow problem make implementation too challenging. 

But identifying resistance isn’t enough. Insurance leaders deal with unknowns all the time. If we all knew exactly what was going to happen with everything, business would be a unicorn. Even when the path forward is uncertain, sitting back and waiting isn’t a strategy.

So rather than just outlining the problems, let me provide a practical roadmap for what you can actually do right now. 

What problem are we solving? Not “we want to use AI” or “our competitors are doing it.” What specific business problem needs solving?

• Who benefits? Is it your customers? Your stakeholders? Your investors? Your employees? Be explicit.
• What’s the ROI case? Figure this out upfront with real specificity. Don’t think you’re going to figure out the ROI later. I see this happen all the time, and it’s a recipe for failed initiatives.

 Here’s the critical part — and I am going to repeat myself— your ROI metrics need to be specific, not generic. “We’re going to improve our loss ratio” isn’t good enough. The whole company is focused on improving loss ratios. You must ask pointedly, “What is this AI solution going to improve? Set up that metric ahead of time with precision.

Make sure your AI strategy embeds into your strategic goals and aligns with the day-to-day pragmatic realities of your operations. If there’s daylight between executive strategy and operational reality, your implementation will fail regardless of how good the technology is.

Here’s a headline: You can’t execute an AI strategy without data.

People want to talk about being “AI-first,” but it’s really got to be “data-first.” We all know that data is messy in every company, but you can’t wait any longer. Figure out how to parse it out. Do some data triage. Identify which parts of your data you need to tackle urgently. 

If data quality, data accessibility, and governance aren’t in your strategic plan, why aren’t they? I know prioritizing data is hard because it’s expensive and time-consuming, but this is urgent right now. The cost of not addressing it is becoming untenable. What is more, you have more tools available now than ever before.

There’s a phrase that captures this moment precisely: “everybody’s a data steward.” What kind of steward are you of your own company’s data?

Everyone needs to answer that question. Data literacy needs to spread across the organization. Do all your employees understand how valuable company data is? Do they treat it like an asset? When they’re generating data—filling in forms, documenting interactions, completing processes—do they have that sense of how important and valuable it is?

Everyone is in the data business now. Whether you’re an agent, in sales, in underwriting, in claims — you’re a data steward. Take it seriously.

For the board:

Ask yourselves: does our executive team have a plan for communicating the value that incorporating AI will deliver? Have they boiled the message down into actionable tasks?

Take some time to spell it out. Boards hear about AI constantly, but they need information distilled into value proposition language that makes sense in the world they live in. As Mark Twain said (more or less): “I apologize that this letter is so long—I didn’t have time to write a short one.”

The board needs to understand how long these initiatives will take to materialize so you can get out of the hype cycle and into sustainable business models. They need to know how you’re dealing with legacy systems, because if you’re not starting from scratch, you’re bolting this on. Explain how you’re doing that in a way that makes sense for your business.

Within the organization:

Adoption is the killer. Keeping in mind that adoption is a very big issue—a topic we covered extensively in my previous article—ask yourself how you can evaluate the readiness of your team.

Who are the internal influencers? Who might derail the initiative? Are you engaging with them early and often?

Now ask yourself about your process. Very few people want AI just rolled out to them. They want to know there was a top-down, bottom-up approach to the strategy. 

Have you aligned strategic goals with the pragmatic realities surrounding how work actually gets done? Answering this means you’ve got to be talking to individual contributors, frontline staff, middle managers.

Have you vetted your strategy and approach? Did you get buy-in by asking people’s opinions? Even if you go a different direction than someone suggested, they’ll appreciate being included rather than feeling like this is being rolled out so their job disappears the next day.

One executive told me something I regard as invaluable: “Rather than us telling our employees how their jobs are going to change, we’ve engaged the employees to help define how their roles will change.”

If 30-40% of your job tasks are going to be done by AI, why not be part of defining how you create more value? This pro-active inclusion translates into measurable results so there’s a clear path forward—not “we’re just going to wipe out all these employees now because we have AI.”

Maybe the sweet spot is, “Our people no longer have to do these menial tasks. We’ve figured out new revenue streams, better conversion, better retention.” Involve employees in helping define what those metrics are and what success looks like.

It’s okay to tell your employees: “With this, we’re going to have to figure out how to add more value, more revenue, more top line and bottom line.” They probably have ideas.

Let me give you some concrete examples of how people can think about AI enhancing rather than replacing their roles.

Sales: Sales is perfect for this. AI tools can handle the grunt work—building prospect lists, vetting contacts, setting up meetings. That’s wonderful. But guess what? Your customers are not going to buy from AI. People still want to buy from people.

As a salesperson, look at your own metrics around your sales process. Can you build your pipeline faster? Qualify better? Could you sign up for a bigger number? Shorten the sales cycle? Increase the value of the sale because you have more time to dig into customer pain points and do the things AI can’t do?

Lean into what improvements you’re going to be able to produce.

Underwriting: Master the data. Master what it’s telling you, what it’s able to glean, what it’s able to process in terms of long documents or complex risk factors. That allows you to hone better risk assessment.

What does that mean for the metrics you’re measured on? Are you able to do better risk selection that fits underwriting guidelines? What improvements can you watch yourself making as you use these tools?

Now own them. Sign up for them. And if the early versions don’t work, get in there and problem-solve with your manager and your team.

In a transparent world where everyone can check everyone else’s work—including AI checking yours—just own it. Be the driver, along with your peers, of the ways you want to be evaluated. Be part of that conversation, too. And as a leader and manager, welcome that.

Maybe what you have is a bad process. Have you done the process engineering to assess system snafus? Should these be fixed first before considering your AI solutions?

In every company everywhere, you run into issues where the system won’t do something, or there’s a hurdle, and it’s not going to be fixed because there’s an IT backlog or whatever. 

So what happens? Workarounds on workarounds. You end up with really clunky, bad processes just because people were trying to figure out a way to get the work done.

Don’t think you can just slap some AI on that.  Legacy issues demand your attention first. Remove the ad hoc bandages, fix the process, and then put AI on it.

If you then have AI automation or AI agents helping with that process, what can you eliminate? 

What should you eliminate?

Use AI to stress-test your strategy

Here’s something powerful that people miss: You can use AI itself to help with your implementation.

Spend time using AI to help with your messaging. Have it be the skeptic and challenger. Have it play devil’s advocate and say: “Where is my message going wrong? Here’s who I’m delivering my message to. Here’s who’s going to be really skeptical. Can you poke at this and tell me what’s going to trigger someone that I didn’t intend?”

If all you’re getting from AI is “Oh, that’s a great point, Kirstin,” you’re not using it the right way. It’s not about an ego boost. Let that AI hurt your feelings now so that you can avoid the worst later. Wouldn’t you rather mess it up with AI than with the whole company? AI is not going to fire you or look at you funny when you walk down the hall.

You can use it as a messaging thought partner to avoid missteps you didn’t intend. People misread texts. They misread emails. This is where AI is your best friend—have it point those things out to you.

This works for employees, and it works for your board. If you’re going to present your ROI and your plan to the board, ask AI to play the role of a board member. Where are they going to tear apart your plan? Use it to get crisp and clear.

Because people are afraid. They’re afraid of: Where is this going to go? Do we have the right data? Is this going to take my job? Are we going to spend a bunch of money and get nothing for it?

That’s actually where AI can really help you address those concerns proactively.

As you grow and scale your AI capabilities, are you thinking about what skills and new competencies you’ll need to add to your company? What kind of new roles? New people? New skill sets?

Are you rethinking how you message to get people who aren’t in insurance to join insurance? Just like with the internet, you needed people from Silicon Valley, technologists who could bridge worlds. It’s the same thing here.

But most of us who have lived through previous digital transformations have the playbook. 

We’ve been through technological transformation before — with the internet, with predictive analytics, with telematics. The companies that succeeded weren’t the ones with the best technology. They were the ones that figured out how to align technology with strategy, get organizational buy-in, fix their data fundamentals, and measure what matters.

AI is both more powerful and more accessible than any previous wave of technology. That’s what makes it urgent. But the principles of successful adoption remain the same: start with the business problem, get your data house in order, bring your people along, and measure results with precision.

Remember: Change is slow until it’s not. And when it accelerates, you want to be ready—not because someone panicked in response to a stock sell-off, but because you did the hard work of building real capability throughout your organization.

AI is already transforming insurance. The question insurance professionals face is whether they’ll be part of shaping that transformation, or let it happen without guiding how it’s deployed.