The Secondary Perils Blind Spot : And the AI Opportunity Hiding in it.

"For two decades, Insurance industry oriented itself around primary catastrophes — named storms, major earthquakes, the once in a decade event that defined CAT management. AI followed that direction. We have built better models, faster claims triage for high-volume surge events, and smarter hurricane corridor pricing. All of it is real and valuable. However, we find them increasingly irrelevant to where losses are actually coming from.

Secondary perils such as hail, flash floods, convective storms, et cetera, now account for over sixty percent of global insured catastrophe losses. These are not rare events that need better models. They are frequent and structurally invisible to AI systems trained on historical CAT data. A hailstorm that generates four thousand claims across twelve zip codes in seventy-two hours does not behave like a hurricane. A flash flood in a neighborhood that has never flooded before does not appear in your ten-year loss history reports. Most AI products are not built for this, and the loss of trends is making that clear.

The good news and the opportunity for us is that the data and the technology to address secondary perils do exist. Real-time weather intelligence has never been more granular. Satellite revisit rates are now measured in hours, not days. Climate science has produced forward-looking physical risk models that make historical triangles look primitive. And generative AI gives us, for the first time, the ability to synthesis across all of these data streams in something close to real time.

What is missing is not capability but the willingness to orient our AI investment around the problem that is actually driving losses today, rather than the problem that was driving losses when we wrote our first Series A pitch deck.

What secondary perils demand is forward-oriented AI — systems that learn from each storm season, operate at parcel level rather than zip code, and feed claims signals back to underwriting within the same season, not the next actuarial cycle.

At Blue Pond, this is exactly the gap we are building into. Not because it is the easiest problem, but because it is the right one. The carriers and MGAs that orient their AI strategy around secondary perils today will not simply improve their combined ratio. They will be the ones still writing property business profitably in 2030 — while others are left explaining, very beautifully, why their reserves were wrong again.