Aipedia: Financial advisers

AI risk profiling and attitude to risk

Aipedia, by the Apex broking team · Last reviewed 2026-10-27

AI and psychometric tools can score a client's attitude to risk, but the adviser must still judge whether the result, and the portfolio built on it, is suitable. A wrong risk band is one of the most common routes into a suitability complaint.

Part of: AI in Financial advisers

In short

Risk profiling tools turn questionnaire answers, and increasingly behavioural or conversational data, into a risk score that maps to a portfolio. AI can make these tools more dynamic but less transparent. COBS 9.2.1R requires the adviser to ensure suitability, which includes capacity for loss, not just willingness to take risk. CP26/10 proposes a single attitude-to-risk concept. Bias, poor handling of vulnerable clients and solely automated decisions raise Consumer Duty, Equality Act and data protection issues. Mis-banding claims usually fall under PI as advice claims.

How advisers use AI for risk profiling

Most advice firms already use psychometric questionnaires that produce a risk score. Newer tools add AI to adapt questions, analyse free-text or conversational answers, combine attitude to risk with capacity for loss and knowledge and experience, and map the result automatically to a model portfolio. Some flag inconsistencies between what a client says and how they answer.

These features can improve profiling, but they also make it harder to explain. A traditional questionnaire produces a score an adviser can trace back to specific answers. An adaptive or behavioural model may not. If the adviser cannot explain why a client scored as they did, it is difficult to show the recommendation was suitable, or to explain it to the client in a way that meets the consumer understanding outcome.

The duties that apply

Suitability: COBS 9.2.1R(1)(a). The score is an input, not the answer; the adviser must test it against the client's circumstances and capacity for loss.
Proposed changes: CP26/10, published 25 March 2026, proposes merging COBS 9 and 9A, a single attitude-to-risk concept and "sufficient" rather than "necessary" information. The policy statement is awaited and the paper does not mention AI.
Fairness: the FCA AI Update maps fairness to the Consumer Duty, Principle 9 (suitability of advice), FG22/5 on the Consumer Duty, the FCA's guidance on vulnerable customers (FG21/1), UK GDPR and the Equality Act 2010.
Automated decisions: under the Data (Use and Access) Act 2025, solely automated significant decisions need safeguards, including human intervention and a way to contest, and special category data is restricted.

What can go wrong

A model trained on general population data may score some groups systematically higher or lower. Clients with low financial confidence, limited English or cognitive decline may answer in ways that do not reflect their real tolerance. An AI tool may weight a single answer heavily, or fail to explain how it reached the score, leaving the adviser unable to justify it. Automatic mapping to a portfolio can hide the step where the adviser should challenge the result.

Illustrative scenario: a retired client completes an AI-adaptive questionnaire on a good day and scores as balanced. The tool maps them to a balanced portfolio and the adviser accepts it. The fact-find shows the investment is most of the client's liquid wealth, so capacity for loss was low. After a market fall the client complains that they were put in the wrong risk band.

How PI usually responds

Mis-banding is a classic suitability allegation, and PI usually responds to it as an advice claim subject to the wording and excess. Insurers commonly look at whether the file shows the adviser considered capacity for loss and challenged the tool's output. A tool that mis-scores a whole segment of clients can create a cluster of claims, which raises aggregation and notification questions. See how PI responds to suitability claims.

Practical controls

Validate the tool before use and ask the vendor how scores are derived and tested for bias.
Test outcomes across client characteristics, as the FCA October 2026 good-practice page recommends.
Assess capacity for loss separately and record how it moderated the score.
Record the adviser's discussion and any override, with reasons.
Offer a non-AI route for vulnerable clients and see the Apex guide to vulnerable customers.
Explain AI use in profiling when telling clients you use AI.

How Apex helps

Apex arranges PI for financial advisers. A broker can review your cover and how your policy treats AI-assisted work, including how insurers view risk profiling processes. See AI in financial advice and AI in cashflow modelling.

Sources

Frequently asked

Does CP26/10 change how I assess attitude to risk?

Not yet. CP26/10 proposes a single attitude-to-risk concept as part of simplifying COBS 9 and 9A, but the policy statement is awaited. Until final rules are published, the existing suitability rules apply. CP26/10 does not mention AI.

Can an AI risk score decide the portfolio automatically?

Automation is not banned, but the adviser remains responsible for suitability. Where a decision is solely automated and significant, the Data (Use and Access) Act 2025 safeguards apply, including human intervention and a way to contest. Most firms keep an adviser review step.

Will PI cover a claim that a client was put in the wrong risk band?

Usually, as a suitability claim, subject to the wording and excess. A well-documented file showing capacity for loss was considered and the score was challenged where needed is the strongest defence.

How do I check an AI profiling tool for bias?

Ask the vendor how the model was built and tested, then run your own checks comparing scores across client groups such as age and sex. The FCA October 2026 good-practice page recommends testing outcomes across consumer characteristics and monitoring live outcomes after deployment.

Check your PI covers how you use AI

Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.

Start a PI quote Request a callback

Apex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.