Having the AI ROI conversation amid the hype

"All the technological changes that have occurred in my time in the industry were slower than we thought," said Duperreault, who has lived through multiple technology transformations. "They produce less than we thought and cost more than we thought."

At the same time, he was clear that leaders do not have the luxury of ignoring AI.

"It's coming," he said. "The one thing you have to do is embrace it."

That tension is the executive challenge.

Move too slowly, and your company risks falling behind. Move too fast, and you may automate broken processes, create governance gaps, damage trust, or burn capital without a clear business return.

Insurance has seen this movie before. The internet was going to eliminate agents. Google and Amazon were going to take over insurance distribution. The insurtech boom was going to rewrite the industry. Some things changed. Progress should be welcomed. But many of the bold proclamations did not materialize.

As Jeff Rieder, Partner and Head of Benchmarking at Aon, told me, "There's always been a lot of hype around new technologies."

He pointed out that while the internet, predictive analytics, online quoting, agency interface tools, and insurtech all changed parts of the operating model. They did not eliminate the core work of underwriting, claims, acquiring customers, and managing the processes around them.

That is the lesson of the hype cycle: The first-order prediction is usually too dramatic, while the long-term change is often more profound than people expect.

Markets are pricing the AI threat

We saw a preview of the market's reaction in February when insurance broker stocks sold off after OpenAI approved an insurance app inside ChatGPT. The app stoked fears that AI agents could disintermediate brokers. Analysts quickly called the selloff overdone, and stocks largely rebounded after the selloff.

The lesson should not be dismissed just because the market overreacted.

Public markets are now willing to punish insurance businesses based on perceived AI exposure before the industry has finished explaining where human value sits.

That creates a communications challenge.

Executives need to explain why AI will change the work of intermediaries, underwriters, claims professionals, and service teams while also clarifying where human judgment, relationships, accountability, and expertise remain essential.

Jeff made this point directly in our conversation about distribution. As customers get more comfortable relying on AI agents, he said there is a real opportunity for the distribution channel to be disrupted, especially in lower-complexity areas. His point was that agents and brokers need to be intentional about how they service customers.

That applies to carriers, MGAs, reinsurers, and service providers.

The operational reality is still hard

The operational reality inside insurance is much harder than the headlines suggest. According to AM Best's April 2026 report, Artificial Intelligence Appears to be Ready, But Most Insurers Are Not, 45% of insurers cited data readiness as a top challenge to implementing AI.

That will surprise no one who has worked inside insurance technology.

The industry is full of legacy systems, acquired platforms, inconsistent data fields, manual workarounds, and institutional knowledge sitting in the heads of people who have been holding the operating model together for years.

Stefan Holzberger, Executive Vice President and Chief Operating Officer at AM Best, described this plainly in our conversation. Most insurers, he said, are "struggling with legacy systems," data quality, the amount of data, and whether their systems communicate with each other.

He also pointed to the customer upside when insurers get it right: the more insurance can leverage data into insight, the faster and more responsive the experience can become for policyholders.

That is the opportunity. Better claims triage and responsiveness. More effective fraud detection. Smarter underwriting and cleaner submissions. More productive employees.

But none of that happens because a company launches an AI pilot.

The investment and impact gap

AM Best found that 66% of survey respondents expect to increase AI investment over the next 12 to 24 months. At the same time, 78% said AI has had no impact on premium growth or retention.

That gap between investment and business impact is where executive credibility is won or lost.

The answer is to talk about AI as operational change and avoid talking about it like a magic layer that spreads across the enterprise.

That means three things.

First, start with the business problem. Are you trying to reduce quote turnaround time? Improve loss ratio? Reduce expense ratio? Increase underwriter capacity? Improve claims cycle time? Reduce leakage? Improve broker responsiveness? Every AI initiative should tie to a business outcome an executive team, board and employees can understand.

Second, be direct about readiness. If your data is fragmented, explain the plan to improve it. If your core systems are limiting speed, show where the constraints sit. If your processes need redesign before automation, make that part of the roadmap. Boards and employees can handle operational reality. What they cannot handle is a vague AI narrative that turns into missed expectations a few quarters later.

Third, explain implementation in plain English. In a boardroom, confusion leads to no. With employees it sows doubt, fear and resistance. Executives need a simple narrative they can champion. And simple explanations take work to get right.

What boards and employees need to hear

The best AI updates answer four simple questions:

1. What are we doing?
2. Why does it matter?
3. How will it change the way work gets done?
4. How will we measure whether it worked?

Rather than: "We are deploying agentic AI to optimize underwriting workflow orchestration."

Try this: "We are using AI to help underwriters sort submissions faster, identify the clean risks earlier, and spend more time on the complex accounts where their judgment matters most. We will measure success by turnaround time, hit ratio, loss performance, broker satisfaction, and underwriter capacity."

That is making the work executable.

AI will not move through insurance at the speed of the headlines. This industry is regulated, complex, relationship-driven, and full of operational debt. Brian's warning still applies: status quo moves as technology matures. What worked yesterday will not necessarily get you where you need to be tomorrow.

The winners will be the companies that can connect AI to business value, operational readiness, talent development, governance, and measurable outcomes.
It's wise to call out what's hype and translate your approach into a plan your people, executives, board, and customers can believe in. Take the time to make it simple and clear. As risk managers, you can mitigate the loss of productivity and performance that happens when people are confused or overly concerned.

PropertyCasualty360 Columnist Kirstin Marr is the Founder and CEO of Lead The Machine, a podcast and newsletter exploring the human side of AI at work. For over 25 years, she has helped more than 100 companies navigate transformation in technology, data analytics, and now AI to modernize and deliver profitable growth. Previously, Kirstin served as Chief Analytics Officer at Insurity and President of Valen Analytics, where she guided analytics businesses from startup through acquisition and scale.

Opinions expressed here are the author's own.

(Featured image credit: Artram/Shutterstock)

Also by Kirstin Marr: Your AI strategy will fail if your culture rejects it