FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Debt advice firms
AI decision tools can rank debt solutions, but CONC 8.3.2R requires advice that is suitable, in the customer's best interests and explains why other options are unsuitable. A tool that steers customers to the wrong solution can affect years of their finances and every case it touches.
Part of: AI in Debt advice firms
In short
Decision tools score a customer's budget and debts and suggest a debt management plan, a debt relief order, an IVA or another route. CONC 8.3.2R requires suitable advice based on a full assessment, with reasons other options are unsuitable. CONC 8.2.2G says recommending a solution the firm knows or ought to suspect is unaffordable is likely to breach the Principles and the Consumer Duty. The FCA has warned about referral fees biasing advice towards IVAs, and CONC now bans debt packager remuneration from CONC 8.3.9R. A biased or poorly tested tool can repeat one error across thousands of cases.
Once a budget and a list of debts are in place, the adviser has to work out which solutions are available and which is suitable. Decision tools, which may use machine learning or generative AI, take those inputs and produce a ranked shortlist, sometimes with a draft rationale. Some firms use them to support advisers; others use them to route customers straight into a particular product journey.
The commercial context matters. Fee-charging firms may earn more from some solutions than others, and the history of the sector shows how incentives can shape advice. The more a tool influences the outcome, the more it needs to be tested for that bias.
Suitability. CONC 8.3.2R requires advice to have regard to the customer's best interests, to be suitable and to rest on a sufficiently full assessment of their financial circumstances. The firm must explain the suitable options and why others are unsuitable. A tool that outputs a single answer without that reasoning does not do the job alone.
Affordability and fairness. CONC 8.2.2G says recommending a solution the firm knows or ought to suspect is unaffordable is likely to breach Principles 2, 6 and 9 and the Consumer Duty, and that discouraging a customer from seeking alternative advice is likely to breach Principle 6.
Incentives. CONC 8.3.6G says firms should not have incentives that undermine CONC 8.3.2R. CONC 8.3.9R to 8.3.17G ban debt packager remuneration. The FCA's 2018 Dear CEO letter, now marked historical, warned that referral fees could bias advice towards IVAs.
Content of the advice. CONC 8.3.4R requires the advantages, disadvantages, costs and risks to be set out in a durable medium, with specific content for IVAs and protected trust deeds (8.3.4R(4)) and DMPs (8.3.4R(3)(e)).
Illustrative scenario: a firm's ranking tool was trained on past cases in which IVAs were recommended more often than the customers' circumstances warranted. It keeps ranking IVAs first for customers with modest debts and no assets, where a debt relief order or a simple plan would be suitable. Advisers usually accept the top result. A past-business review later finds a pattern of unsuitable recommendations.
The ways this goes wrong are familiar:
Where an IVA follows, insolvency practitioners face their own duties at onboarding; see AI in IVA and debt solution onboarding.
A claim by a customer that negligent advice put them into an unsuitable IVA or DMP is usually the sort of civil liability PI is designed for, subject to the terms and excess. A systemic problem raises harder questions: aggregation of related claims under one limit, and whether costs of an FCA-required redress exercise are covered at all. Fines are commonly excluded. Because PI for debt advice is voluntary, these points are entirely about the wording. See does PI cover AI errors in debt advice and aggregation of claims.
Apex arranges PI for debt advice firms. A broker can explain how your policy deals with a series of related claims from one tool. See the AI in debt advice hub.
The rules do not ban it, but CONC 8.3.2R still requires suitable advice based on a full assessment and an explanation of why other options are unsuitable. Keeping the adviser as decision maker and using the tool to support them is the safer model.
Remove commercial factors from the logic, test outcomes by solution type and customer group, and compare recommendations with a sample of adviser-led cases. Keep the results, because the FCA expects auditable records and firms remain responsible for outsourced AI.
That can lead to many complaints and possibly a past-business review. How PI responds depends on the wording, including aggregation and whether regulator-led redress costs are covered. Tell your insurer early if you identify a pattern.
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 callbackApex 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.
