SECTOR INTELLIGENCE BRIEFING, BASED ON DOCUMENTED PUBLIC EVIDENCE, RESEARCH DOCUMENT
01The Documented Picture
◈ This briefing draws exclusively on documented, publicly available evidence from UK regulators, parliamentary committees, and authoritative surveys. The predictive findings in Section 04 are inferences from this evidence base, not assertions of findings BreakPoint has made in this sector.
75%
of UK financial services firms using AI
BoE/FCA Survey, Nov 2024
55%
of AI use cases include automated decision-making
BoE/FCA Survey, Nov 2024
46%
have only partial understanding of AI they use
BoE/FCA Survey, Nov 2024
0
published studies measuring adviser AI override rates in UK WM
Research gap: documented absence
AI is embedded across the UK wealth management sector in three primary roles: client risk profiling and capacity-for-loss assessment, model portfolio construction and automated rebalancing, and suitability letter drafting. In each case, the governance framework describes the AI as "advisory" or "decision-support" and states that advisers retain final discretion. The FCA's 2024 AI Update confirms Consumer Duty as the primary regulatory lens for assessing these deployments.
The critical evidential gap: there are no published studies, no regulatory findings, and no sector-wide data on adviser AI override rates in UK wealth management. The absence of this measurement is itself a governance signal. A claim that advisers retain discretion that is not substantiated by override rate data is a claim that is asserted but unevidenced. This is the central tension BreakPoint is designed to resolve.
02The Governance Gap, What the Evidence Shows
The Treasury Select Committee's January 2026 report explicitly called for the FCA to publish, by end of 2026, guidance on "the level of assurance expected from senior managers under the SMCR for harm caused through the use of AI." This recommendation reflects the Committee's finding that the current framework, which assigns AI accountability to existing SMFs without specifying what decision-level exercise of that accountability looks like, is insufficient. In wealth management, where advisers certify suitability and senior managers certify oversight, the Committee's finding applies directly: having an accountable person named is not the same as evidenced accountability for AI-shaped outcomes.
The FCA launched an AI Live Testing programme in 2025, with the first cohort exploring AI use cases including targeted support, Open Finance-driven personalisation, and digital/non-advised mortgage channels. The FCA's own testing programme, which explicitly acknowledges that AI is reshaping the boundary between advice and non-advice, is evidence that the sector's AI deployment is materially ahead of the governance and regulatory clarity available to govern it. Firms deploying AI in suitability-adjacent roles are operating in a space where even the regulator is still actively testing what appropriate governance looks like.
46% of firms have only a "partial understanding" of the AI they use, primarily because a third of all AI use cases are third-party implementations. In wealth management, where risk profiling AI and model portfolio construction tools are frequently supplied by third-party platform providers, advisers are often exercising "discretion" within the output of a system whose internal logic their firm only partially understands. Discretion that operates within a framework the adviser cannot interrogate is constrained discretion, and the constraint is invisible unless measured.
Under FCA Consumer Duty, firms must demonstrate that advice and products deliver good outcomes for clients. The FCA has confirmed that this applies where AI shapes the suitability process, including risk profiling, model portfolio matching, and suitability letter generation. The duty is outcomes-based: firms cannot satisfy it by pointing to a governance framework that describes adviser discretion. They must be able to evidence that the advice clients received reflected genuine human judgement about their specific circumstances. No firm has yet been required by the FCA to demonstrate this at decision level from operational records. That situation is unlikely to persist.
MiFID II Article 25 (retained in UK law) requires firms to obtain information about the client's knowledge, experience, financial situation, and investment objectives sufficient to make a genuine recommendation. Where AI risk profiling shapes the information-gathering stage, determining which risk band applies before the adviser sees the client, the suitability basis for subsequent advice is built on an AI determination, not an independent human assessment. The FCA has not yet published specific guidance on this gap but the structure of the regulation does not accommodate AI pre-determination of suitability basis without a named human challenge authority at the profiling stage.
03The Regulatory Framework
FCA Consumer Duty (PS22/9)
Suitability advice must deliver good outcomes, demonstrated from operational evidence. AI-shaped suitability assessments are in scope. Adviser discretion must be real, not assumed.
Risk profiling, model portfolio matching, suitability recommendation, suitability letters. Any AI-shaped stage of the advice journey that affects the client outcome.
MiFID II Article 25 (UK)
Genuine suitability assessment requires real information and real human judgement. AI pre-determination of risk band or product set constrains the suitability basis without a named challenge authority at that stage.
AI risk profiling at intake. AI model portfolio matching. Any AI that determines which options the adviser presents to the client before the adviser applies independent judgement.
FCA SM&CR
Senior managers certify accountability for AI-active processes. Treasury Select Committee has called for FCA guidance on what decision-level accountability for AI harm looks like in practice, guidance not yet published.
Head of Advice, CRO, COO with AI-active suitability processes in scope. Framework certification without decision-level exercise is the documented gap.
FCA COBS 9A, Suitability
Firms must take reasonable steps to ensure advice is suitable. Where AI shapes the information on which suitability is assessed, "reasonable steps" must include named authority over AI-shaped suitability inputs.
Risk questionnaire AI. Capacity-for-loss AI. Model portfolio suitability matching. Letter generation AI where content touches regulated advice substance.
04What BreakPoint Would Predict
Epistemic note: The following are indicative and predictive inferences from the documented evidence base above - not assertions of findings BreakPoint has made in this sector. Each prediction identifies a specific gap type that the evidence strongly suggests is present. BreakPoint's role is to determine, for a specific organisation, whether the gap exists and its severity. Note that BreakPoint is not limited to these factors however, and will often surface unexpected risks or risks that were not fully articulated.
Every wealth management firm operating AI-assisted suitability recommendations describes the AI as "advisory" and advisers as retaining discretion. No firm has published its adviser AI override rate. The documented absence of this measurement, combined with the BoE/FCA survey finding that 55% of AI use cases include automated decision-making and 46% of firms have only partial understanding of their AI, is strong evidence that the override rate is low and unknown. BreakPoint would measure it. A low measured override rate does not constitute a breach; it constitutes an obligation to explain it, and to restructure the workflow if the explanation is that advisers are ratifying rather than deciding.
Evidence basis: BoE/FCA Survey 2024; FCA Consumer Duty (outcome evidence required); MiFID II Article 25 (genuine suitability assessment); research gap, no published override rate data in UK WM sector.
AI risk profiling determines the risk band within which the adviser constructs the suitability recommendation. It is the foundational AI-shaped input to the advice process. The evidence strongly suggests that no named individual holds documented authority to challenge or override the profiling output at the point of its production, before it shapes which options the adviser considers. The adviser's discretion operates within the band the AI has established. That is not independent professional judgement, it is constrained professional judgement, and the constraint is AI-set.
Evidence basis: MiFID II Article 25 (information-gathering stage); FCA COBS 9A (reasonable steps for suitability); BoE/FCA Survey 2024 (46% partial understanding, particularly third-party profiling tools); Treasury Select Committee 2026 (SMCR accountability guidance absent).
Every governance framework in the sector asserts that advisers retain discretion. What the frameworks do not specify, because they cannot specify it from documentation, is whether the workflow sequence places AI consultation before or after independent adviser assessment. The documented evidence from the BoE/FCA survey (55% automated decision-making; 46% partial understanding) is consistent with a sector-wide pattern of AI-first workflow architecture that governance documents describe as adviser-led. BreakPoint tests the sequence, not the framework description.
Evidence basis: BoE/FCA Survey 2024; FCA Consumer Duty (outcome evidence vs. documented intent); FCA AI Update 2024 (no AI-specific rules, existing frameworks apply, which means existing gaps also apply).
AI suitability letter drafting tools are approved for administrative and structural content. The FCA's AI Live Testing programme explicitly includes exploration of where AI can support advice-adjacent content, which signals that the advice/non-advice boundary is live and contested in this area. BreakPoint would test whether the current scope of suitability letter AI extends beyond the approved scope at time of governance sign-off, and whether the review process applied to letters matches the actual content being generated.
Evidence basis: FCA AI Live Testing 2025 (targeted support and advice-adjacent AI in scope for regulatory exploration); FCA Consumer Duty (communications must support informed decisions); FCA COBS 9A.
Automated model portfolio rebalancing operates between Investment Committee review cycles, often weekly or event-triggered. The Investment Committee reviews the model ranges; it does not hold operational authority over individual rebalancing signals. Between reviews, AI-triggered rebalancing operates without a named human challenge authority at decision level. The governance framework describes Investment Committee oversight; the operational reality is AI execution within approved parameters, without a named human whose authority covers the period between reviews.
Evidence basis: BoE/FCA Survey 2024 (automated decision-making in investment processes); PRA SS1/23 (model risk management, independent challenge required, not periodic committee review); MiFID II Best Execution.
05Source References
BoE/FCA 2024
Bank of England / FCA,
Artificial Intelligence in UK Financial Services 2024, November 2024.
bankofengland.co.uk →Treasury Committee 2026
House of Commons Treasury Select Committee,
Artificial Intelligence in Financial Services, January 2026. Recommendations include FCA SMCR guidance for AI harm and AI stress-testing.
parliament.uk →FCA Consumer Duty
FCA, Consumer Duty Final Rules and Guidance, PS22/9, July 2022. Effective July 2023. Applies to AI-shaped advice outcomes per FCA AI Update 2024.
MiFID II Art 25
MiFID II Article 25, Suitability assessment requirements. Retained in UK law. Requires genuine suitability assessment based on adequate client information. FCA COBS 9A implements this domestically.
FCA AI Live Testing
FCA AI Live Testing programme, launched 2025. First cohort includes targeted support, Open Finance personalisation, digital advice channels. Confirms advice/AI boundary is live regulatory territory.
fca.org.uk →FCA AI Update 2024
FCA,
AI Update, 2024. No AI-specific rules; Consumer Duty and SM&CR are primary frameworks.
fca.org.uk →
Luminary Diagnostics · BreakPoint: AI Control Gap Assessment
Every wealth management governance framework says advisers retain discretion.
None of them measure whether that discretion is exercised.
BreakPoint measures it, from operational records, not policy documents. Override rate, workflow sequence, risk profiling authority. The determination a firm can act on and defend.
Book a BreakPoint conversation →