70 to 80x Faster: Inside CFC’s Blueprint for Agentic AI Underwriting

Most insurers are still asking what AI might do for them one day. CFC has already answered the question: its agentic underwriting system is live in production, cutting a week’s worth of complex cyber submissions down to an hour. In the latest episode of The Voice of Insurance podcast, host Mark Geoghegan sits down with George Beattie, Head of Innovation, and Chris Mullan, Head of Data and AI at CFC, for a conversation that starts with eye-opening productivity numbers and ends somewhere far more provocative: a case for why AI exclusions will not survive contact with reality, and why the entire architecture of the insurance product is up for grabs.

70 to 80x: What Live Agentic Underwriting Looks Like

CFC’s agentic system is already handling live quotes in its cyber book, and the numbers are hard to ignore. “It’s 70 to 80 times,” Mullan tells Geoghegan, describing the gain in inbox-to-quote speed. Where a complex submission once took a week to work through, an underwriter and an agent can now clear it in an hour. The headline stat, though, is the turnaround: “We’re seeing inbox to quote… under 15 minutes for quite complex risks,” against a previous benchmark where only 70 to 80% of submissions were answered within 24 hours.

What makes this section of the conversation worth hearing in full is the philosophy behind the numbers. Beattie is blunt about where he thinks the market is going wrong: “I think the point of AI is to increase the humanity in the system, not to reduce it.” Flip the maths on a broker who currently spends 70% of their time re-keying data, he argues, and that time goes straight back into the conversations that build the business – the relationships, the cross-sell, the trust that no agent can replicate.

Beyond Faster Horses: Redesigning the Product, Not Just the Process

Beattie’s sharpest warning is aimed at insurers using AI purely to accelerate what already exists. “If all AI does is perpetuate the existing business model, are we missing the point of what this thing allows us to do,” he asks, “which is to rethink and redesign critical elements of our infrastructure and how we work?” His reference point is Henry Ford’s apocryphal line about customers who, before the car, would only ever have asked for faster horses.

Applied to product design, that means confronting how little sense some packaged coverage makes once you look closely. SME package policies, Beattie points out, ship with built-in redundancy: buy one section of a six-section policy and you still receive all six, whether you need them or not. Mullan’s analogy for the alternative is a good one: rather than starting with a block of marble and chiselling it down, LLM orchestration lets insurers build a policy the way additive manufacturing builds a part – tailored, piece by piece. “You just build what you need when you need it,” Mullan says. It is a genuinely different way of thinking about how a policy gets assembled, and one that CFC believes removes the administrative burden that has made true tailoring impractical until now.

Why AI Exclusions Are Heading for an Embarrassing Climb-Down

The most striking part of the episode is Beattie’s case against treating AI as a separate, excludable peril. His view has shifted markedly from a few years ago, and it is built on a simple idea: AI is becoming as fundamental to business as electricity or water, which makes it almost impossible to isolate as a distinct cause of loss. “AI is going to be impossible to derive from normal business in the future, which means it’s like a utility,” he explains.

That has direct consequences for wording. Rather than bolting on second-order AI-specific products, CFC is building affirmative AI coverage into its existing D&O, cyber, and other core lines, so the policies explicitly respond when AI is a cause of loss instead of leaving the question ambiguous. Beattie is unambiguous about the alternative: “It’s going to be quite an embarrassing climb-down for those that are using broad scale exclusions.” He draws a direct line to earlier moments the market would rather forget, when computer-use exclusions were quietly abandoned once they proved unworkable, arguing that “past history has shown the insurance market needs to be very careful about reacting from an exclusionary perspective in the face of change.”

The Full Picture

Productivity gains measured in multiples, not percentages; a genuine rethink of how insurance products are built; and a considered, well-researched argument against the industry’s instinct to exclude first and ask questions later – Beattie and Mullan cover a huge amount of ground in this episode, and this is only part of the conversation. They also discuss how agent-to-agent distribution might reshape parts of the market, why specialty insurance still captures a vanishingly small share of global GDP, and what genuinely new products AI could make viable for the first time. For anyone weighing up what AI means beyond the next efficiency gain, this is a conversation worth hearing in full.