UK consumer insurance in 2030

The FCA has published the Mills Review, Sheldon Mills' independent assessment of how AI could transform retail financial services by 2030. It deserves a close read from everyone in insurance, because its predictions land harder here than almost anywhere else.

What the Review predicts for consumer insurance

The central shift the Review describes is from financial services that are human-led and episodic to services that are AI-enabled, continuous and delegated. It maps the journey as an autonomy spectrum: AI that first explains, then recommends, then acts with the customer's approval, and eventually acts continuously within limits the customer has set.

For insurance specifically, it predicts cover moving closer to the point of need: embedded distribution, automated quote comparison, claims guidance, and agents that shop the market at renewal. Stakeholders told the Review that this "hyper-switching" would be a pro-competitive force across insurance, and the demand signals are already real: one in five UK adults say they would use AI that acts autonomously within pre-set goals, and around nine million already put financial questions to general-purpose AI apps.

It also carries a warning. As agents take over search and comparison, the AI-mediated interface becomes a new source of market power. Intermediation does not disappear, it relocates: away from brokers and comparison sites, towards agents, platforms and the data layers beneath them. Only around 30% of home and 20% of motor policies are sold direct today. Where those journeys begin in 2030 is now an open question for every insurer and every brand that distributes cover.

Why this matters most in insurance

Insurance is where the problems the Review wants AI to fix are most entrenched. Only 30% of UK adults hold protection policies (such as life or income protection). Just 28% of renters have contents insurance. Loyal customers routinely over-pay through inertia, and most people engage with their cover once a year at best. An agent that understands a customer's circumstances, explains what cover actually does, and acts at the right moment addresses each of these directly.

But the Review is equally clear about what stands in the way: trust. Three-quarters of AI users trust AI to explain things clearly; only around one in five trust it to handle their personal information responsibly. Trust, control and access, it concludes, will decide adoption. And it makes a striking observation for firms: a reputation for trusted AI processes could win business. In an AI-enabled market, governance is not overhead, it is a growth strategy.

How we get there

From our own work deploying consumer-facing AI in regulated insurance markets, we believe the path to 2030 rests on three foundations.

Safety you can evidence. Every AI conversation should be protected in real time, as well as evaluated after the fact, with every answer traceable to the policy language behind it. Not because regulation demands it, though it will, but because delegation is only rational when the delegate can show its working.

Products that machines can read. The Review notes that firms will need to present products in ways that are legible to AI systems as well as humans. Agent-led comparison is only as good as the product data underneath it. Whole-of-market cover data, structured, current and machine-readable, is the infrastructure the agentic era depends on.

Autonomy earned in stages. The right way through the autonomy spectrum is sequential. Start with AI that explains and guides, factually and without steering. Earn the right to act with customer approval. Only then move to continuous optimisation within limits the customer sets, with consent, mandates and audit built in from the first stage.

The Review opens with a question: who is AI going to serve? In consumer insurance the answer has to be customers, and the firms that serve them best in 2030 will be the ones building the trust infrastructure now.

Building for 2030, today

At Open, this is the future we have been building towards. Our platform powers embedded insurance for some of the world's biggest brands, and AI Companion powered experiences already guide customers through quote and buy journeys, protected by AI Safety Harness and grounded in structured, machine-readable product data. Every layer is designed so that, as customers delegate more, the trust, evidence and control the Review calls for are already in place.

If you are an insurer or a brand working out what the Mills Review means for your customers, get in touch to work with us or to find out more about what we are building.

You can read the full Mills Review on the FCA website.