Executive summary
UK payments and retail firms report high levels of AI inclusion readiness, but their reported practices and the experiences of customers with characteristics of vulnerability reveal a significant evidence gap.
Confidence is not evidence
A survey of 110 senior decision-makers at UK firms deploying consumer-facing AI produces an inclusion readiness score of 4.1 out of 5, placing respondents in the “ready” band. However, firms’ confidence in their readiness is consistently stronger than the practices they report. This gap is evident across all three dimensions of the index and widens when firms assess their future readiness.
A third of firms do not monitor outcomes for vulnerable customers as a distinct cohort, limiting their ability to demonstrate whether AI-enabled services are delivering comparable outcomes. A parallel survey of 2,000 UK adults reinforces this concern: 17% of vulnerable adults said they had been unable to complete a banking or payment task in the past year because a tool did not work for them, 3.1 times the rate among other adults. Taken together, the findings suggest that firms’ confidence in AI inclusion is running ahead of the evidence available to support it.
Key findings
Firms are confident they are ready
Firms give themselves an average AI inclusion readiness score of 4.1 out of 5, with little difference between payments firms and merchants.
But reported practice does not always match
Across every area measured, firms rate their readiness more highly than their reported practices suggest. The gap is widest when firms assess their readiness for the future.
One in three cannot track outcomes for vulnerable customers
While 65% of firms monitor outcomes for vulnerable customers separately, the remaining 35% rely on overall performance or investigate only when problems arise.
Support varies by form of vulnerability
Specific provision is more common for some forms of vulnerability than others. Firms report less specific provision for customers experiencing recent adverse life events.
Vulnerable customers are three times more likely to encounter problems
17% of vulnerable adults report being unable to complete a banking or payment task because a tool did not work for them, compared with 5% of other adults.
Firms want greater regulatory clarity
Almost half of respondents cite regulatory clarity and guidance as the biggest barrier to making AI-enabled journeys work effectively for vulnerable customers.
Implications for the payments industry
AI has significant potential to improve access and outcomes across payments, but the findings suggest that deployment is advancing faster than firms’ ability to demonstrate inclusive outcomes. Closing that gap will require better measurement and more consistent provision for vulnerable customers, rather than slower adoption. Firms that can evidence customer outcomes will also be better placed to demonstrate that their AI-enabled services are meeting regulatory expectations.
Firms
Measure the cohort before defending the outcome
Segmenting completion, abandonment, escalation and complaint rates by the FCA's four drivers is a reporting cut rather than a redesign, and it is what turns a Consumer Duty assertion into a Consumer Duty defence.
The payments sector
Design for vulnerability that arrives, not vulnerability that is declared
With 26.4 million UK adults inside the FCA's definition in May 2024, most of them transiently and without telling their provider, provision built around disclosure will miss the people who need it.
Digital verification providers
Treat inclusion as the product, not the phase after it
Since the government cancelled the national digital ID scheme in July 2026 while continuing GOV.UK Wallet, the inclusion duty sits with the payments and financial services sector rather than with a government programme.
Regulators
Publish the worked examples the sector is asking for
The FCA's commitment to publish AI good and poor practice before the end of 2026 answers the barrier firms name most, and that publication will only be as strong as the practice it can point to.
Methodology
The industry survey (UK30577) covers 110 senior decision-makers at UK firms deploying consumer-facing AI: 60 in financial services, reported here as payments, and 50 in retail and wholesale, reported as merchants. It is unweighted throughout, and all percentages are raw counts over the base. Within financial services the sub-types are fintech 30, bank 21 and payment service provider 9; sub-sector, role and AI-involvement cuts are directional only and none leads a finding. Consumer figures come from the parallel survey fielded by The Payments Association's D&I sub-working group (UK29904), of 2,000 UK adults, weighted and nationally representative, four months earlier. The two are set side by side as supply and demand, not compared statistically. Fieldwork August 2026 and April 2026 respectively.