InsurTech Geek Podcast

There are genuine nuggets - parametric BI mechanics, the cloud-monitoring tech stack, data center SLA product structure, and specific loss ratios - but roughly half the episode is personal backstory, conference small talk, and an unfocused AI tangent that generates no actionable insight for operators.

“We basically run every service in every Availability zone in every region for those three cloud providers all day, every day”

“our cloud product is probably running, uh, around 20. Like all of our other products are Fairly new, last 18 months”

Applying parametric triggers to cloud downtime and data center SLAs in the commercial lines space is a genuinely uncommon idea with real structural novelty, but the back third of the episode collapses into well-worn AI-threat talking points that add nothing fresh.

“we were basically running cloud services and gathering data and built this cloud downtime product. It was basically a way to bring parametric insurance from the property market into the commercial insurance space”

“there's a contractual liability exclusion. You know, how do they, how does that cover SLAs, right”

Rick Wong is a legitimate practitioner with a coherent 20-year arc from AIG ops to Hiscox underwriting to building an insurtech product line from scratch - real operator credentials - but he is Head of Insurance at a small early-stage startup, not a scale-tested C-suite executive, which caps the caliber ceiling.

“I was the first underwriter, only underwriter. And then as that grew, we hired like three or four people in New York”

“we have one data center account where we're trying to build a $500 million tower”

Above-average for the format: named loss ratios (~20 on cloud product), per-risk limits ($20 - 150M, one $500M tower), product launch months, a four-hour waiting period, 25-day claims cycle, Q1 GWP at 60% of full-year 2025, and a named US competitor (Descartes). The AI speculation section, however, is entirely evidence-free.

“we finished Q1 60 of what we did in all 2025. We're projecting to be. Over 200 growth this year”

“we're putting up anywhere from like 20 to 150 million. Yeah. Per risk”

The host asks reasonable product questions but routinely hijacks the conversation with his own anecdotes (daughter at NYU, his coding history, his data center experience), never pushes back on unverified growth claims or pricing logic, and allows the final third to drift into undirected AI speculation that the guest has no special authority to address.

“I've been writing code since I was 11, 19, uh, 91, 92, 93, 94. I started write code before the Internet and we would hack our friends computers”

“That's pretty amazing”