Company secretaries supporting boards of regulated firms are increasingly using AI to draft board-pack summaries of Consumer Duty annual board reports, validate Principle 12 mapping for committee minutes, prepare director briefings on PRIN 2A obligations, and reconcile FCA Feedback Statements such as FS25/2 against existing supervisory expectations recorded in board papers. The output sits at the centre of director-attestation packs and audit-committee minutes that auditors and the regulator can request.
Two frontier AI models tested by the RLB Specialist Panel produced 9 substantive failures on this regulation under audit conditions. The failure classes recorded are: Misstated Statutory Architecture, Inference Drift on the Foreseeable-Harm Safe Harbour, Confused Guidance with Rule on Consumer Testing, Inference Drift on Fair Value Quantification Expectation, Hedge in Place of Verified FS25/2 Figure, Refusal to Confirm a Documented FS25/2 Count, Reversed the PRIN 2A Group-Insurance Exclusion, Invented Dual-Event Timeline for a Single FS25/2 Withdrawal, Refusal to Confirm FS25/2 Withdrawal Count.
Questions were prepared by the RLB Specialist Panel based on real practical AI usage in the workflows the respective audience uses AI for, and each finding is bound to verbatim regulator-issued source text held as primary substrate. The Consumer Duty (PS22/9 introducing Principle 12 and PRIN 2A, in force for open products from 31 July 2023 and for closed products from 31 July 2024) is the central retail-conduct regime the FCA now uses to grade firm behaviour, and the failure modes seen here all land inside the day-to-day work product that company secretaries sign off on.
For company secretaries, the operational consequence is direct. A board-pack summary or audit-committee briefing built on the AI's framing imports a defect into director attestations. The next supervisory visit, an internal-audit pull of the board record, or an external review of governance materials will surface the gap, and the secretariat carries the governance-quality exposure.
Citation IDs for the findings in this brief: RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q002-Sonnet46, RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q003-Opus47, RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q007-Sonnet46, RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q008-Opus47, RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q013-Opus47, RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q013-Sonnet46, RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q018-Opus47, RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q020-Opus47, RLB-H-GB-FCA-CONSUMER-DUTY-PS22-9-Q020-Sonnet46. Each citation links to the per-finding record, the AI subject answer, and the regulator-issued substrate excerpt the answer was tested against. The RLB Specialist Panel maintains an audit-traceable record of which model produced which answer, against which substrate passage, and the binding is what makes the finding referenceable in firm work product and in supervisory correspondence.
The findings below are the ones that company secretaries working under the Consumer Duty are most likely to encounter in the AI tools they already use, and the briefing sections that follow read each finding against the regulator-issued text.
This is the consolidated view of findings. Click the Citation IDs or 'see details →' on any item for the full details for each finding.
Company secretaries drafting Consumer Duty board-paper authority sections need a clean statutory citation. A board paper that cites only FSMA 2000 misses the post-Brexit FSMA 2023 framework, which leaves the directors unable to engage with the full architecture in the rare event the question is escalated. The cost is not direct enforcement, but the loss of board-paper authority that flows from an incomplete legal underpinning.
Company secretaries supporting board oversight of Consumer Duty compliance need accurate accounts of the cross-cutting rules' actual structure. The model's multi-condition safe harbour is the kind of detail that will be included in a board-pack 'rule summary' and become the firm's working understanding. A director's challenge that surfaces the discrepancy with PRIN 2A.2 puts the secretary in the position of explaining why a board paper imported a fabricated test.
Company secretaries preparing the annual Consumer Duty board report under PRIN 2A.8.3R need to distinguish rule-level from guidance-level obligations correctly. The model's conflation of FG22/5 with PRIN 2A.5.10R is exactly the kind of error a non-executive director's challenge would surface, and the company secretary's role is to ensure the board report does not carry it forward.
Company secretaries reviewing the annual Consumer Duty board report's fair-value section need to ensure the methodology described in the paper matches the FCA's actual expectation. If the AI's affirmative quantification framing reaches the board pack, the board approves a methodology that exceeds what FG22/5 requires, and the firm carries an unnecessary methodological commitment that will be hard to unwind.
Company secretaries supporting the firm's regulatory-history reporting need accurate accounts of supervisory communications. The model's fabricated April and August 2025 timeline, if it reaches a board paper on supervisory expectations, will undermine the paper's authority on review. A simple FS25/2 citation is the correct fact; the AI's invention is not.
Company secretaries summarising the FCA's supervisory-letter landscape for the board cannot use the model's evasive answer. The board expects a concrete account; an evasive AI response leaves the secretary without a basis for the paper, and a downloaded FS25/2 will resolve the question in a single read.
Company secretaries preparing scope memos for the board on Consumer Duty applicability across the firm's distribution chain need the carve-out preserved. The AI's reversal of the group insurance exclusion, if imported into a board paper on Duty scope, leads the directors to approve a Duty-compliant programme across activities the FCA has excluded.
Company secretaries supporting board-level engagement with the FCA's published supervisory record need accurate accounts of which letters and reports remain in force. The model's repeated fabricated timeline across multiple questions is a strong signal that AI-assisted summaries of this content cannot be trusted without source-document verification.
Company secretaries reviewing AI-assisted briefings need to treat fabricated supporting citations as a structural failure mode. The Clifford Chance citation in the model's evasion response is exactly the kind of plausible-looking source that escapes a quick review but fails any deeper check, and the board secretary's role includes catching these before they reach the board pack.
Every finding on this page compares an AI subject's account of the rule against the regulator's verbatim text from the regulator's own portal. Both are linked. Each delta, its root causes, and impact analysis are documented and published with immutable Citation IDs.