Executive summary
Findings highlight reported financial crime exposure, demand for practical AI governance guidance and questions about investment in intelligence sharing.
This report combines FC360 live polls, registration responses, the Pulse Survey and interviews. The Pulse Survey found that 92% of respondents reported at least one of ten listed financial crime risks in the previous 12 months. Event polling identified AI regulation as the change with the largest difference between selections for greatest uncertainty and best preparedness. These findings describe respondents’ experiences and judgements; they are not independent assessments of firms’ controls.
Key findings
The risk landscape
92% of Pulse respondents reported at least one of ten listed financial crime risks in the previous 12 months.
Priorities for 2026 and 2027
Fraud prevention was the biggest challenge for 24% of registration respondents; AML and transaction monitoring effectiveness followed at 20%..
The gap
AI regulation showed a 17-percentage-point difference towards uncertainty among 59 respondents answering both Fraud Strategy questions.
AI-enabled fraud
Across two session polls, 56% of responses selected AI-enabled fraud as the risk increasing fastest relative to response capability.
Intelligence sharing
In the Fraud Strategy poll, 63% selected real-time sharing as the leading factor in the national strategy’s success.
What members should do next
Document AI control ownership, review intelligence-sharing arrangements and identify specific questions requiring regulatory clarification.
Industry implications
Members can act before new guidance or legislation arrives. First, record the AI systems used in fraud and anti-money laundering (AML) controls, their owners, approval arrangements and monitoring. Second, review which fraud signals the firm shares, with whom, through which channels and under what arrangements.
The surveys identify concerns and investment intentions. They do not measure the cost of governance approaches, the effectiveness of controls or the speed of intelligence flows. The recommendations that follow are proposals informed by the findings, rather than outcomes proved by them.
Banks
Document AI control ownership and oversight
For AI used in fraud and AML controls, record the business purpose, accountable owner, approval process, monitoring and escalation arrangements. Include systems supplied by third parties. Report gaps and remediation dates to the responsible executive committee.
Payment service providers
Prepare data for common sharing standards
Map the fraud signals your firm sends and receives, the sharing channels it uses and the formats each requires. Identify gaps in data quality, coverage and response times. Assess potential participation in Pay.UK’s proposed Enhanced Data Exchange (EDEx), alongside existing channels, using current service specifications and participation requirements.
Technology platforms and telecoms providers
Share actionable signals before funds move
Work with payment firms to identify fraud signals that can be shared lawfully and acted on before funds move. Agree contact points, escalation routes, data-quality requirements and feedback on whether a signal helped prevent harm. Broader sharing obligations and accountability should be identified separately as policy proposals.
Regulators and policymakers
Clarify practical expectations and disputed-payment scenarios
TPA recommends practical examples for AI in fraud and AML controls, clarification of disputed AI-initiated payments, and a clear approach to broader fraud intelligence sharing. Existing rules continue to apply. These requests seek greater clarity and coordination; they do not imply that firms can defer current responsibilities.
Methodology
The report uses four evidence sources with different participants, questions and collection methods. Results are descriptive and should not be treated as representative of the payments industry. Differences between sources are not measures of change over time or direct measures of a gap between attitudes and behaviour. Response counts vary by question.
- FC360 live polls: 16 single-choice Slido polls across eight sessions, with 135 participants reported overall. Individual questions had different response counts. Results from fewer than 30 responses are illustrative; results from fewer than 15 responses are reported as counts rather than percentages. Participation was voluntary.
- Registration responses: 2,034 registrations were recorded, and 1,063 people attended. Financial crime questions received responses from 582 people, with response counts varying by question. Respondents were self-selected and included people from different sectors and roles.
- Pulse Survey: Opinium surveyed 100 senior UK financial crime decision-makers for The Payments Association between 30 June and 20 July 2026. Findings describe this sample; subgroup comparisons should be read cautiously, particularly where fewer than 30 respondents are involved.
- Interviews and contributed quotations: The report includes filmed interviews with Kamlesh Harry, Anna Sweeney, Humera Yakub, Simon Miller and Erez Nounou, and quotations from Chris Ainsley, David Crawford, Azy Shojaeian and Humzah Amin. Interview takeaways are editorial summaries, separate from direct quotations.
Regulatory references describe the position at publication in November 2026. Survey and poll results relate to the collection periods stated above. Recommendations and proposed deadlines are the report’s suggested actions, rather than commitments announced by government or regulators.