Eight in 10 UK consumers now use an AI assistant at least three times a week, and six in 10 use one to help make decisions - a shift that is already reshaping how people choose financial products. Research presented at Defaqto's 2026 Data of Record conference found that more people now turn to AI-powered tools than to a provider or financial adviser when making those choices.
The numbers sit alongside a trust deficit that has not kept pace with adoption. One in five consumers actively distrust AI tools when choosing a financial product. Nearly one in five have already acted, or almost acted, on AI guidance that later turned out to be wrong.
Most striking is that 62% of consumers say they cannot judge whether AI guidance is accurate. A client who cannot evaluate that guidance cannot self-correct when it fails. The research, carried out by Savanta on behalf of Defaqto, suggests a substantial share of the market is making product decisions in exactly that position.
Mike Piddock, managing director and interim CEO at Defacqto, told delegates that greater reliance on AI increases rather than reduces the importance of data quality. Firms need confidence in where data has come from, how it has been structured, and whether it gives a broad enough view of the market.
Defaqto has a commercial interest in that argument - its Star Ratings business depends on independently verified data. But the point carries more weight beyond the conference room than it might at first appear: AI outputs are only as reliable as the information they draw on, and consumers have no direct way to assess either.
Alex Ward-Booth, director at Savanta, compared the challenge to the California Gold Rush. Success there depended on distinguishing genuine gold from fool's gold, and those who could not tell the difference lost their stake.
Keynote speaker Emma Boardwell, founder of Emotional Finance, argued that firms will need to "zig towards speed, efficiency, automation and AI in order to compete" while also "zagging towards human connection in order to differentiate and matter."
That framing maps onto what insurance-specific research has been showing. The 2026 Guidewire European Insurance Consumer Survey found that only 30% of UK consumers are comfortable with AI making decisions about their insurance policy price. The three conditions consumers attach to any AI acceptance in insurance are human intervention, transparency, and independent regulation.
Those conditions describe what brokers already provide. A broker-assisted client has a named professional accountable for the advice, access to a broad market view, and a human to call when something goes wrong. A growing share of clients who turn to AI instead are getting none of those things, and many discover that only after the fact.
The Chartered Insurance Institute's July 2024 Public Trust Index found that broker-assisted consumers report better insurance outcomes than those using price comparison sites or going direct, including on claims speed. As AI tools become a more common first stop for insurance decisions, that gap in outcomes is likely to widen before it narrows.
