Reimagining Risk in the Age of AI

with Amir Kabir from Overlook Ventures

Amir Kabir has spent the last decade doing what a lot of people in insurance say they want to do: figuring out where the real opportunity is. As Founding & Managing Partner of Overlook Ventures — and before that one of the first three people on the founding team at Munich Re Ventures — he’s backed a generation of insurtechs including Next Insurance, Slice, HDVI, and Bought By Many.

In this week’s conversation, Amir joined hosts James Benham and Rob Galbraith to unpack what the insurtech era got right, what it got wrong, and where the next wave of opportunity is already forming. It’s one of the sharpest reads of the last ten years you’ll hear from someone who had a seat at the table the entire time.

All Episodes Available On

Technology cycles. Risk doesn’t.

The whole Overlook investment thesis sits on one sentence: “Technology is always going to cycles, but risk always evolves and emerges through those cycles.”

Blockchain, cloud, mobile, AI — every tech wave has a peak and a trough. Risk is the permanent category underneath. That’s why Overlook doesn’t invest in tech categories. It invests in companies building what Amir calls the infrastructure of risk: simulation software for autonomy, foundation models for robotics, cyber incident response, deepfake detection, new insurance products for hardware.

“Technology moves in cycles. Risk always evolves through them.”
— Amir Kabir, Overlook Ventures

Insurance is a get-rich-slowly business

If there’s one line from this episode that’s going to travel, it’s this one. Amir uses it to reframe the entire last decade of insurtech.

The early thesis — roughly 2014 through 2018 — was that insurance was an outdated, customer-hostile industry ripe for SaaS-style disruption. Build a better UX, wrap an MGA in tech, and scale the way consumer startups scale.

It didn’t work the way people thought. Not because the observations were wrong (incumbents really are slow; customer experience really is poor) but because insurance isn’t a scalable SaaS business. It’s a risk business. Underwriting mistakes made at the front end cost you at the back end. Regulated, saturated, price-driven, and highly competitive markets don’t bend just because you have a prettier app.

“Insurance is a get-rich-slowly business.”
— Amir Kabir, Overlook Ventures

The real takeaway for investors: a $10B exit in pure insurance is probably off the table. A $1B outcome is not. And the tech-enabled MGAs that worked were the ones that picked niches incumbents genuinely couldn’t serve — Next Insurance for small businesses, Insure (formerly Hippo-adjacent) for taxi and black-car drivers, Bought By Many for pet insurance — rather than trying to disintermediate carriers head-on.

What AI actually changes (and what it doesn’t)

James’s framing on AI is worth restating: it’s the most disruptive tool he’s seen in 25+ years of shipping software for insurance. For the first time in his career, “no, we can’t do that” has become “yes, but we need to do some automation work first.”

Where AI is real and already working:

Claims automation — ingesting unstructured data, predicting outcomes, reasoning across form sets
Underwriting data extraction — pulling structured signal out of historically unstructured inputs
Software plays adjacent to admitted lines — tooling for brokers, agents, and carrier ops teams

Where AI still runs into the regulator:

New admitted products on AI alone — you can go E&S, but you still have a distribution problem
Fully autonomous underwriting in saturated consumer lines — the 50-state regulatory surface hasn’t moved

Amir’s summary of his own portfolio is a useful frame: roughly a third of Overlook looks like traditional insurtech. The other two-thirds are enterprise software plays that happen to work across insurance, financial services, and other regulated markets — because building a $10B outcome selling only into insurance is a very hard timeline to pencil.