The Voice of Insurance: The discipline to walk away in a more competitive market

Some podcasts are essentially a victory lap. This one wasn’t.

It was more practical than that. A discussion about focus, underwriting discipline, and what you don’t do when the market starts getting friendlier again.

The guest was Richard Milner, CEO of Chaucer Group. Chaucer isn’t a start-up and it isn’t trying to pretend it is. It’s a business with a long memory, and Milner talks like someone who has seen enough cycles to know where “easy growth” usually ends up.

No reinvention story here

Milner isn’t selling a transformation narrative. There’s no grand “new Chaucer” pitch. The direction is simpler and, frankly, harder to execute.

Do more of what Chaucer is genuinely good at. Make peace with the things it isn’t. Then act like you mean it.

That means leaning into the specialist lines where the group already has underwriting depth and market standing, and being prepared to trim back in areas where it doesn’t have a clear advantage. The underlying point is one most people agree with in theory and fewer follow in practice: you don’t get paid for being present in every class. You get paid for being good in the ones you choose.

Lloyd’s matters, but so does choice of paper

One of the more grounded parts of the conversation is how Chaucer thinks about platforms.

Lloyd’s is core to the identity and the trading model. But it isn’t always the neatest fit for every client, every structure, every jurisdiction. That’s where the non‑Lloyd’s platform comes in.

Chaucer’s Dublin company has become a sizeable part of the group, and the point isn’t to “move away” from Lloyd’s. It’s to widen the tool kit. Sometimes Lloyd’s paper is exactly what a client needs. Sometimes it isn’t. If you’re trying to be useful to global clients, having options matters.

Milner describes the two platforms as complementary rather than competing. One doesn’t cancel out the other. It gives the business more ways to put a deal together without forcing everything through a single route.

Competition is back. That’s when the mistakes start

Milner is direct about market conditions. Things are getting more competitive. You can feel it in property cat. You can feel it in the way underwriters are being tested on renewals.

Casualty is a different conversation, but the theme is the same. If you’re not prepared to walk away when the maths don’t work, you end up storing up pain. And the pain rarely arrives immediately. It arrives later, quietly, while everyone is congratulating themselves on growth.

What Milner keeps circling back to is this: premium targets are not a strategy. If you let the target do the driving, you’ll make bad decisions when the market gets soft and you’ll be explaining them for years.

Delegated authority is part of the furniture now

On MGAs and delegated authority, Milner isn’t dramatic about it. He doesn’t talk as if it’s a novelty or a threat. He talks as if it’s simply part of how specialty insurance now operates.

There are obvious reasons for that. Distribution is changing. Underwriting talent is expensive and finite. Delegated models, done well, can be efficient and scalable.

But his view isn’t “more delegated authority at any cost”. It’s more selective.

Strong underwriting franchises will do well and keep growing. Some won’t. Some will fall away. That isn’t a shock. It’s what happens when the market gets more competitive and the easy tailwinds disappear.

The key point is governance. Delegated authority can work brilliantly, but only if it’s genuinely underwritten and properly controlled. Delegated does not mean hands‑off.

Algorithms help with speed. They don’t replace judgement

Chaucer has been more measured than some peers on algorithmic underwriting, and Milner sounds comfortable with that pace.

He frames it as a strategic choice rather than reluctance. If you’re going to bring automation into underwriting, you need to know where it fits, what it improves, and what it can break. Moving fast is not impressive if it creates blind spots you find later in claims.

Milner’s instinct is that underwriting still has a human core. Relationships still matter. Service still matters. The market still remembers who behaves well when things are messy.

You can automate parts of the process. You can speed up decisions. You can reduce friction. But there are moments in insurance where judgement is the product, and an algorithm doesn’t carry reputational risk. People do.

AI: useful in operations, awkward as an exposure

Milner is interested in AI, but not in the way people sometimes talk about it at conferences.

He’s not selling it as magic. More like a tool that can remove drudge work, sharpen decision‑making, and help a business operate more efficiently.

The more interesting part is the risk angle. AI isn’t only something insurers use. It’s also something clients are building into operations, products, and decision systems, and that will show up in claims eventually. The market is still working out what “AI exposure” even means in practical terms, and where it sits across cyber, tech E&O, PI and beyond.

Milner sounds cautious, but not closed. The tone is: we need to understand what the loss looks like before we pretend we can price it.

The unglamorous bits: claims and talent

Two topics come up that often get treated like footnotes, even though they decide reputations.

First, claims. In specialty, claims isn’t back office. It’s where clients decide whether you’re a partner or just a logo.

Second, talent. Milner’s point is basically that the market can’t keep solving its people problem by poaching. If you want technical depth in five years, you need to bring people in earlier, train them properly, and keep them learning.

If you don’t, you end up with a thin “middle”, expensive hiring, and a lot of institutional knowledge walking out the door.

The point of the plan

This isn’t a flashy strategy. It won’t win an awards dinner.

But it’s the kind of plan that tends to survive a cycle: focus on the lines where you have genuine edge, be realistic about where you don’t, use the right platform for the right job, and don’t let a softer market talk you into decisions you’ll regret.

That’s the thread running through the whole episode. Not headline-chasing. Just doing the job properly.