A reported case in which a chatbot told a mortgage borrower that a payment holiday would not affect their credit score highlights a central risk in adviser AI adoption: plausible-sounding answers can still be wrong. Payment arrangements, credit-file impacts and lender treatment vary by circumstance, product and firm policy.
AI can be valuable for meeting notes, document summaries, client-service routing and identifying missing information. It should not be allowed to give unsupervised, personalised mortgage, credit or debt guidance. This is particularly important as the FCA warns consumers about pressure tactics, misleading information and unsuitable debt solutions.
Practical controls:
- Require human approval for client-facing answers on credit, arrears, affordability and product suitability.
- Ground internal AI tools in approved, dated lender and compliance content.
- Maintain audit trails, test outputs regularly and provide a clear escalation route to an adviser.