The Consumer Duty asks UK lenders to prove good outcomes with data, which is where AI agents help: testing affordability, reviewing arrears and forbearance handling, comparing outcomes across customer groups and drafting board report evidence. Compliance and senior managers keep every judgement. The FCA's consumer credit rules, SM&CR and UK GDPR also apply.
What does UK consumer lending compliance spend its time on?
Affordability testing comes first. FCA rules require a creditworthiness assessment that considers whether the borrower can repay without significant adverse effect on their finances. Compliance samples approvals and checks that income and expenditure were properly considered, that the lender did not simply rely on a score, and that the decision fits the evidence.
Borrowers in difficulty are the second focus. Lenders must treat customers in arrears or financial difficulty with forbearance and due consideration, and identify those who are vulnerable. Compliance reviews contact records, payment arrangements and fees charged during arrears.
The Consumer Duty frames everything. Lenders must deliver good outcomes for retail customers across products and services, price and value, consumer understanding and consumer support, monitor those outcomes with data, and report to the board at least annually. Complaints, ombudsman cases and remediation programmes take the rest.
Where can AI agents help?
- Affordability file testing (depth 1). The agent reads the application, open banking or bank statement data, and the lender's assessment, and flags cases where declared expenditure, verified data and the decision diverge.
- Arrears handling review (depth 1). The agent reviews contact records, arrangements and charges for customers in arrears, and flags fees applied during forbearance, missed vulnerability signals and pressure in calls.
- Outcome monitoring by cohort (depth 2). The agent compares outcomes across customer groups, such as arrears rates, complaints and early repayment, and drafts commentary on differences for compliance. Our piece on whether an AI agent can monitor customer outcomes for Consumer Duty explains the method.
- Support journey testing (depth 1). The agent tests how easily customers can get help, change arrangements or complain, including where an AI support agent is the first point of contact. See Consumer Duty and AI support agents.
- Board report drafting (depth 2). The agent drafts the evidence sections of the annual Consumer Duty board report from monitoring results.
Outcome monitoring by cohort is where the Duty makes an agent most useful. Showing whether customers with characteristics of vulnerability fare worse requires analysing large volumes of data regularly, which an agent can do and a small team cannot.
What stays with people?
The board, and the senior manager responsible for the Duty, own the conclusion on whether customers receive good outcomes. The agent drafts evidence; people decide what it means and what changes.
Remediation stays human. When testing finds unaffordable lending or unfair arrears fees, deciding which customers are affected and what redress they get is a judgement with financial and regulatory consequences.
Individual vulnerability decisions stay with trained staff. An agent can flag that a customer mentioned illness or bereavement; a person decides how to adapt treatment. Our view of AI compliance agents as second line, never the last word is that the agent widens what compliance sees and people decide.
Which UK rules apply?
The UK takes a principles-based, regulator-led approach to AI rather than a single AI law. The FCA regulates consumer credit, including creditworthiness assessment, treatment of customers in arrears and financial difficulty, and the Consumer Duty, which requires good outcomes for retail customers across the four outcome areas and an annual board assessment. SM&CR assigns named senior managers to these areas, including a board champion for the Duty. Our piece on FCA Consumer Duty and AI agents in lending goes deeper on the Duty specifically.
Banks are also supervised prudentially by the PRA, whose SS1/23 sets model risk principles that reach credit models. UK GDPR and the Data Protection Act 2018, enforced by the ICO, restrict decisions based solely on automated processing with significant effects, such as refusing credit, and require transparency about how decisions are made. Our overview of how AI agents are regulated in the United Kingdom covers the rest.
This is an explanation of the rules, not legal advice; confirm how they apply to you with counsel.
How do you start in the first 90 days?
Month one: run arrears handling review at depth 1 across the last quarter. It tends to find the most urgent harm, and the results feed straight into Duty monitoring.
Month two: add affordability file testing on a sample of approvals, with compliance confirming each flag. Map each use to the responsible senior manager.
Month three: pilot outcome monitoring by cohort at depth 2, using it to draft one section of the next board report. Agree with the board champion how the evidence will be presented.
Questions people ask
Can an AI agent write the annual Consumer Duty board report?
It can draft the evidence sections, but the conclusions and the board's assessment belong to people. The board must be able to show it reviewed and challenged the content.
Does UK GDPR stop lenders using AI to decide credit?
It restricts decisions based solely on automated processing with significant effects unless a condition applies, and then requires safeguards including human intervention and the right to contest.
How does the UK compare with the US for consumer lending compliance?
The US focuses on fair lending and specific reasons for credit denial rather than an outcomes duty. See our brief on compliance agents in US consumer lending.