Sponsored Content
Let’s imagine a mortgage adviser, Fiona, on a Friday afternoon. Her current customer is self-employed with a variable income, a historic credit default from what seems like a lifetime ago, and a short-term future filled with worry of the unknown. Fiona listens, understands and gives the customer a list of their options. This is where she feels she is best positioned to help.
Yet too often, what Fiona gets is a queue of reports to pull and requests to the data team that will likely take several days. Fiona is illustrative, but many advisers will recognize her week.
Demand on advisers is rising
UK Finance reported in December 2025 that around 1.8 million fixed rate mortgages were due to end in 2026, many that were agreed far lower than today’s rate. Their mid-year health check showed that refinancing was broadly on track, with £184 billion completed by the end of June, a little over half of the £338 billion forecast. This pressure is set to continue with the Bank of England projecting just over five million households will see repayments rise on refinancing by the end of 2028.
That pressure lands on a profession that is not currently growing. As of 30 June 2026, 35,047 people held permissions to advise on or arrange mortgages, up less than 1% from the end of 2025. Meanwhile, the Intermediary Mortgage Lenders Association’s (IMLA) latest tracker shows advisers placing an average of 105 cases a year in Q2 of 2026, up 9% from Q1 and the highest since 2010. More customers are coming through, but the number of qualified professionals is not growing at the same rate.
Additionally, regulation is moving towards more tailored engagement. Since 2025, dialogue between a firm and a customer no longer automatically triggers advice, and the FCA’s CP26/18 proposals include a tailored interactive dialogue where standard evidence of repayment is hard to obtain. Good conversations are needed at a time when advisers are the scarcest.
Where AI comes in
The AI debate often swings between “it will replace advisers” and “it won’t change anything”. Feedback to the FCA shows that neither is likely, with respondents agreeing that qualified advisers should review and sign off AI-based recommendations. This aligns to what consumers value when making some of the biggest financial decisions of their lives, a person who understands their situation and stands behind their advice.
Instead of the debate, lenders are better considering where AI would give advisers time back. Much of an adviser’s day to day work is not what a customer sees. They might be looking for information, chasing management information, and checking positions. Remove those activities and the human part of their job gets bigger and more meaningful.
Asking questions in plain English
One of the most practical changes that can be made is conversational. Fiona could ask, ‘Which customers have a fixed rate ending in the next three months?’, ‘Which applications are waiting on documents?’ or ‘What has changed for this customer since their last review?’. Instead of requesting a report and waiting or running an analysis herself, she can get an answer in seconds.
Three things make conversational AI trustworthy; the data must be accurate and accessible, it must understand mortgage context, and it must be explainable. This enables advisers and compliance teams alike to understand how an answer was reached. Without all three of these elements, AI tooling becomes a risk instead of help.
The ideal route is in taking a staged approach. Start by asking questions of trusted data, then add further automation only where it is low risk and appropriate with a person always reviewing the outcome
From reporting to conversation
Management information usually means a request to a specialist, a wait, and an answer that arrives after the conversation has moved on. Plain-language questions turn reporting into something advisers and managers can do themselves, without manual data extraction or reconciliation.
This also helps leaders see what might need attention. Questions can be saved as a dashboard or run recurring checks, so the system surfaces issues rather than waiting for someone to identify them. For staff monitoring customer outcomes, this means earlier sight of problems.
What advisers and lenders should ask
Before adopting conversational AI, ask four questions. Is the data ready? Does the AI understand mortgage context? Can you see how it reached an answer? And is a human kept in the loop?
The AI tooling should never minimise an adviser’s judgement but instead provide them with more time with customers.
SBS AI is built around these principles, with reporting and plain-language questions at its core with underlying products kept deterministic and auditable. SBS has been working with UK mortgage providers to enable AI solutions across core systems. To find out more, visit the SBS website.