Aipedia: Mortgage brokers
A mortgage broker is responsible for what its chatbot tells borrowers, including anything that amounts to a personal recommendation. A bot that answers "which mortgage should I get" can cross into regulated advice without the safeguards or records the rules require.
Part of: AI in Mortgage brokers
In short
Website and messaging chatbots help brokers pre-qualify borrowers, answer questions and book appointments. Generative AI chatbots can go further and suggest lenders, products or borrowing amounts. That can amount to a personal recommendation without a fact-find, suitability assessment or record, and can make claims that breach the financial promotion rules. The Consumer Duty requires firms to support customer understanding and avoid foreseeable harm. If a bot pre-qualifies people out, data protection safeguards for automated decisions may apply. PI may respond to bot-generated advice if the wording treats it as the firm's professional services.
Typical uses include answering questions about the process, estimating how much someone might borrow, collecting details for a callback and booking appointments. Messaging bots handle the same tasks over chat apps. Generative AI chatbots answer free-text questions, which means they can be asked anything, including "should I fix for two or five years" or "which lender will accept my credit history". Questions about lenders and criteria are where bots can overreach, because they draw on the same kind of data as AI sourcing tools without an adviser checking the result.
Giving general information about mortgages is not the same as recommending a particular mortgage to a particular person. A bot that takes someone's circumstances and says a specific product or lender suits them is moving towards a personal recommendation, without the fact-find, suitability assessment and records that an advised sale requires. FS25/6 notes respondents' concern about accountability gaps and asks for qualified advisers to review and sign off AI-based recommendations (paras 4.24 to 4.26). A bot that recommends directly has no such sign-off.
Bots can also make statements that would breach the financial promotion rules if written on a web page: overstating rates, implying guaranteed approval or omitting key risks.
Illustrative scenario: a borrower tells a broker's website chatbot about their income, a recent default and their deposit, and asks which lender to use. The bot names a lender and says they "should be accepted". The borrower makes an offer on a property, pays for a survey and applies, and the lender declines. The borrower complains that the broker told them they would be accepted. There is no fact-find, no adviser involvement and only a chat log.
Other problems include a bot quoting a rate that has been withdrawn, telling a borrower they cannot get a mortgage when a specialist lender would help, and storing conversations with sensitive data on a third-party platform.
If a bot's answer is treated as negligent advice, the borrower is likely to claim against the firm. Whether PI responds may depend on whether the policy defines professional services broadly enough to include automated output from the firm's own channels. Some policies may be silent on AI, which can help or hinder. A policy meeting MIPRU 3.2.4R must cover Ombudsman awards. Data exposure through the bot is more commonly a cyber matter. See does broker PI cover AI mistakes and AI exclusions and clauses.
Apex arranges PI for mortgage brokers. A broker can review whether your policy treats chatbot output as part of your professional services. See the mortgage broker AI hub for more.
A bot can give general information, but recommending a specific mortgage to a specific person is regulated advice that requires a suitability assessment and records. Brokers can limit bots to information and route advice questions to qualified advisers.
It may be, depending on what it says and to whom. Statements about rates, approval chances or products can amount to promotions, so bot wording should be checked against the financial promotion rules as with any other marketing.
A clear statement that the bot does not give advice helps, but it may not be enough if the bot answers advice questions in detail. Regulators and the Ombudsman may look at what actually happened rather than the small print.
Only if the bot makes a solely automated decision with a legal or similarly significant effect. If it does, for example by excluding someone from advice, UK GDPR Art. 22C safeguards apply, including human intervention and the right to contest.
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.