Making Risk Flow | The Future of Insurance

The episode contains a few genuinely non-obvious ideas — notably that codifying underwriting expertise reveals practitioners perform better than they believe they do, and that broker indexation is the structural key to facilities becoming permanent. However, roughly half the runtime is hedged generalities, filler language, and standard softening-market observations that any attentive London market participant already holds.

“what people think they do when they underwrite and what you codify doesn't always end up being what they actually do. And actually what they actually do is better.”

“if you're investing for niche lead and delivering capacity lead, you're putting your effort in the wrong place”

The lead/follow spectrum taxonomy (true lead → niche lead → capacity lead → portfolio follower → capacity follower) is a useful codification, and the point about AI shifting technology architecture from process-aligned to decision-aligned is moderately fresh. The AI commentary is otherwise generic ('hallucinations are much less bad,' 'still in POC territory'), and the market-cycle and softening observations are well-trodden.

“the technology architectures will shift to being process aligned, to being decision and capability aligned, much more so. And that is a very different architecture from today.”

“There is a lot out there where this myth of expertise is a myth.”

Greg Brown is a credible insurance-specialist strategy consultant who has done substantive LMA-commissioned research, giving him genuine market access. However, he is an advisor rather than an operator who has underwritten, built a syndicate, or run a P&L — his perspective is synthesised from client work rather than first-hand execution, which limits the practitioner authority on offer.

“we're a strategy advisory firm. We focus only on insurance”

“We've done a lot of work with clients on codifying underwriting rules”

The episode earns credit for naming specific market players and structures — Key's pure-follow syndicate spun out of Brit, Beazley Smart Tracker, QBE QPS, and Canopus — and for a concrete 30–40% facilities market-share estimate and the '90/10 in a simple spreadsheet' codification anecdote. What is missing is any hard performance data, loss ratios, growth rates, or financial outcomes that would validate the claims.

“you've got Beasley Smart tracker that have really invested over the last few years growing that business separate syndicate, you've got the QBE QPS QBE portfolio solutions. We've got Canopus”

“people are talking about 30, 40%. If you add up all of the different types of facilities”

The host makes a genuine attempt to challenge the guest — pushing back on the automation binary and on whether followers need more underwriting rigour when behind a capacity lead — and the guest productively disagrees. Most questions, however, are long, multi-part, and self-answering, and the host repeatedly telegraphs the expected answer ('I think your answer is going to be, it's not that black and white'), reducing productive tension.

“I don't entirely agree with your premise that if you're a follower following a capacity lead, you'd necessarily want to do more underwriting”

“I would love to get your thoughts, because obviously I think your answer is going to be, it's not that black and white”