AI Hallucination ResearchAudiencesSectorsUnited KingdomFinancial Advisory › Compliance
Financial Advisory × Compliance — United Kingdom · published 2026-05-26 · methodology v2.1

AI Hallucinations Affecting Compliance at Financial Advisory Firms in the United Kingdom

Findings — impact summary

This is the consolidated view of findings. Click 'see details →' on any item for the full details for each finding.

  1. Finding 1. Legal basis of Consumer Duty — FSMA 2023 disclaimerRLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q002

    A Compliance team drafting an internal briefing or regulatory mapping note on the legal basis of Consumer Duty may circulate a document that does not address the FSMA 2023 position — leaving the firm unable to respond accurately if the FCA or an external auditor raises this point directly. While the practical regulatory risk of this specific omission is lower than some other findings, it undermines the firm's credibility in regulatory dialogue and creates a gap in its documented understanding of the Duty's foundations. The FCA can draw adverse inferences from a firm that demonstrates incomplete command of the legislation underpinning its obligations.

    see details →
  2. Finding 2. Foreseeable harm carve-out — the "reasonably believes" qualifierRLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q003

    A Compliance team relying on the AI's version of the foreseeable harm carve-out would build a more burdensome and inaccurate standard into its internal procedures — requiring advisers to satisfy additional conditions (good faith, supported understanding, avoiding own-conduct harm) that are not part of the rule before the carve-out applies. More critically, by replacing the firm's 'reasonable belief' test with an objective customer-understanding test, the AI's answer changes the nature of the standard in a way that can affect both how cases are assessed internally and how the firm responds to an FCA supervisory query about a specific customer outcome. The FCA can impose remediation requirements and financial penalties where it finds a firm has systematically misapplied Consumer Duty rules.

    see details →
  3. Finding 3. Retail customer definition — micro-enterprises and charitiesRLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q005

    A scope analysis or client-classification procedure built on the AI's description of the retail customer definition may incorrectly exclude charities that fall within the £1 million annual turnover threshold (if the firm uses the wrong measure) or fail to flag the sourcebook variation (leading to inconsistent application across different business lines). For a Financial Advisory firm with charity clients near the threshold, misclassifying a retail customer as a professional client removes the full Consumer Duty protection — a position the FCA can challenge in a supervisory review and that may require retrospective remediation of the services provided to that client.

    see details →
  4. Finding 4. Consumer testing of communications — rule versus guidanceRLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q007

    A Compliance team that treats consumer testing as a binding rule obligation — rather than a guidance recommendation — may impose mandatory testing requirements on communications production that the FCA does not actually require, creating disproportionate operational burden. More significantly, if the firm's communications policy documents consumer testing as a rule requirement, this creates a compliance gap if testing is missed: the firm has self-imposed a standard it then fails to meet, which the FCA may treat as an internal policy breach even though the underlying regulatory obligation did not require it. The distinction between PRIN 2A.5 rules and FG22/5 guidance also matters for how the FCA can enforce: guidance non-compliance is handled differently from rule breaches.

    see details →
  5. Finding 5. Fair value assessment — quantification of non-monetary benefitsRLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q008

    A fair value assessment methodology designed around the AI's characterisation of the FCA's expectations — treating quantification of non-monetary benefits as encouraged or required — would impose unnecessary analytical obligations on the firm and potentially generate internal pressure to produce quantified figures for items the FCA accepts can be handled qualitatively. Where the firm then presents its fair value assessment to the FCA in a supervisory context, a methodology misaligned with the FCA's stated expectations may invite questions about whether the firm's approach is proportionate or whether the qualitative assessment it did produce was robust. Conversely, a firm that builds its methodology on the AI's inverted account may feel it has satisfied the FCA's requirements when it has not addressed the qualitative assessment the FCA actually expects.

    see details →
  6. Finding 6. FCA withdrawal of pre-Consumer Duty Dear CEO letters — FS25/2RLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q013

    A Compliance team that holds incorrect information about the FCA's Dear CEO letter withdrawal programme — including fabricated dates and a misunderstanding of which document governs the withdrawal — may fail to update its supervisory correspondence register accurately, treat withdrawn letters as still live expectations, or brief the business incorrectly on the current supervisory landscape. Where the firm retains policies or procedures anchored to Dear CEO letters that are no longer in force, it may be applying outdated supervisory expectations alongside current Consumer Duty obligations, creating internal inconsistency. The FCA can form a view about a firm's supervisory awareness from how accurately it tracks changes to the regulatory framework.

    see details →
  7. Finding 7. FCA senior official statements on first-year Consumer Duty complianceRLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q016

    A Compliance team that bases its understanding of the FCA's first-year Consumer Duty concerns on an unverified AI summary of senior official speeches risks constructing a picture of regulatory priorities that may not accurately reflect what the FCA has said publicly. Where that picture is used to prioritise internal Consumer Duty review work, allocate Compliance resource, or frame board or senior management reporting on regulatory risk, the firm may be directing effort toward topics that are not the FCA's stated focus and underweighting topics that are. The FCA's public statements about supervisory priorities carry significant weight in determining where enforcement and supervisory activity will concentrate.

    see details →
  8. Finding 8. Changes between CP21/36 consultation and PS22/9 final rulesRLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q017

    A Compliance team using AI-generated accounts of CP21/36-to-PS22/9 differences to document the evolution of its Consumer Duty framework — for example in a board-level implementation review, a regulatory change management record, or external audit workpapers — would be relying on an account of regulatory history that cannot be verified against primary sources. If specific wording changes are mischaracterised, the firm's record of why it interpreted particular provisions in certain ways may be inaccurate. Where the FCA or an external auditor asks the firm to demonstrate how it tracked and implemented the final rules, a documented account based on unverified AI reconstruction is a weak foundation.

    see details →
  9. Finding 9. Dear CEO letters remaining in force after Consumer Duty implementationRLB-F-GB-FCA-CONSUMER-DUTY-PS22-9-Q020

    A Compliance team that holds fabricated or absent information about which Dear CEO letters remain in force may maintain policies and procedures that reference withdrawn letters as live supervisory expectations, creating a compliance framework anchored to requirements that no longer exist in that form. Equally, where the firm is uncertain which letters remain in force, it may apply an over-broad set of supervisory expectations without a clear basis. The FCA's FS25/2 statement is the definitive reference for this question; a firm that cannot demonstrate awareness of it may face questions in supervisory engagement about how it tracks changes to the FCA's supervisory framework.

    see details →
← Other sector case studies in United Kingdom The detailed Case study →