Over the past year, I’ve found myself living in two insurance worlds at one time. In one, I am deep in the fundamentals at UW-Madison learning the mechanics of risk, the precision of underwriting, and the rigid logic of the balance sheet. In the other, I’m in meetings with founders and mentors who are looking to fundamentally change how those very things work.
From that vantage point, it’s hard to miss how the insurtech story has shifted. We have moved past the era of trying to replace the industry and entered what many call Insurtech 3.0, a wave defined by companies that make the existing system work better.
To get there, the market had to go through a reset. A few years ago, insurtech felt unstoppable. Capital was easy, valuations were rich, and you could raise money on the promise of being the next industry darling with a slick front end and grand expectations. Then reality showed up. Loss experience was bumpier than the pitch decks depicted; reinsurance got tighter, and investors started asking basic questions about combined ratios and cash burn. Funding slowed and the tone around “disruption” cooled off quickly. What’s left now is a smaller, more sober market that cares a lot more about whether the economics actually work.
What often gets lost in that narrative is that the workflow and infrastructure side of insurtech never went away. Even during the peak hype years, there were teams quietly focused on submissions, pricing engines, claims tools, and broker workflows. They just didn’t fit the new carrier vs. old carrier storyline, so they stayed in the background.
Today, those same kinds of companies sit at the center of the conversation. When people talk about insurtech in 2025, they are usually talking about enablers. These are software and services that help carriers and MGAs execute the strategies they already have, not startups trying to replace them. This 3.0 era is less about who owns the customer and more about who owns the data and the workflow.
AI is a big part of why that enabler story has become reality. It serves as an amplifier for human intelligence, helping underwriters and brokers iterate through concepts quickly and experiment with new ideas in real time. This synergy between human intuition and machine speed allows us to move past routine data entry and think more strategically about the entire insurance value chain. Once a carrier has to answer what they are doing with AI, the discussion becomes clear. You end up mapping where data actually lives, who touches it, and where work gets stuck. That exercise shines a light on operational choke points that have been there for decades.
The startups finding traction now pick one painful workflow, learn it in detail, and plug into existing systems instead of trying to rip them out. They focus on unstructured data, the messy PDFs and handwritten notes that have historically trapped risk management professionals in manual entry. By automating the data soup, they move numbers that show up on a P&L.
For a UW-Madison risk and insurance audience, this third act is interesting because it lines up closely with how we are trained to think. We are taught that insurance is a business of inches where a 1 percent improvement in the expense ratio is a massive win. Insurtech 3.0 lives in these details: combined ratios, loss adjustment expense, and distribution efficiency. Their success depends on whether they can make underwriters and actuaries more effective, not on whether they can convince the industry it needs a whole new brand.
This doesn’t mean the first wave was pointless. The early digital carriers forced incumbents to take data and experimentation more seriously. They attracted talent into the industry that might never have considered insurance otherwise. Many of those people now sit inside carriers or second-generation startups, applying what they learned in a more grounded way.
When someone asks whether insurtech is “back,” I don’t look at funding charts. I think about how many conversations now start with a specific operational problem and a realistic role for technology in fixing it. That is what an enabler-driven third act looks like. It’s quieter than the first wave, but for those of us planning to work in and around this industry, it feels closer to the version of insurtech that was supposed to exist all along.
Categories: