Artificial intelligence (AI) is increasingly influencing the UK mortgage market and the broader home-buying process. Use cases range from automating document processing to generating mortgage recommendations and, in some pilots, digitising entire property transactions.
Examples and pilots
Daniel Hegarty, formerly chief executive of Habito, described how the Anthropic Claude AI platform produced detailed, structured advice, including sensitivity analysis on interest-rate movements and comparisons of fixed versus variable-rate products. A Lloyds Bank survey in 2025 of 5,000 people found that more than one in five respondents had already used AI for mortgage advice.
Today the most common application is automated interpretation and standardisation of documents submitted by customers — for example income certificates and bank statements — which speeds up processing and reduces human error. The next step is AI-generated specific recommendations for brokers or customers; these suggestions are currently reviewed by trained advisers, although some scenarios envision AI eventually performing much of the advisory role.
Risks and limitations
There are significant pitfalls. AI-generated recommendations can be based on outdated or incorrect information: one broker reported a client requesting a product that no bank actually offered after relying on ChatGPT’s suggestion. Lenders’ underwriting rules contain many small, practice-based nuances that brokers learn through experience and which are hard to capture automatically.
AI also struggles with sensitive, life-event information such as plans for starting a family or potential career changes, which can materially affect credit assessments. From a regulatory perspective, the lack of explainability in AI decisions is a major concern — banks must be able to show how they reached particular conclusions, which is difficult with “black-box” systems.
Ongoing developments
Despite these challenges, major lenders are experimenting. Lloyds planned a test phase in the second half of 2025 for an AI-based chatbot aimed at intermediaries to answer complex mortgage application questions and speed up underwriting. NatWest, working with the Australian digital platform Pexa, completed a transaction that refinanced a mortgage in just two days. Nationwide and the UK local land registry took steps in February to enable digital signing of mortgage deeds.
The Financial Conduct Authority (FCA) has also involved additional firm groups, including Barclays, Lloyds and Experian, in live testing of AI applications. The long-term aim is full digitisation of the home-buying process: currently it takes on average 104 days from offer to exchange, blamed in part on antiquated legal procedures and paper-based workflows.
As part of digitisation efforts, the UK tech firm Coadjute is developing a blockchain-based registry to link banks, estate agents, solicitors and mortgage advisers directly; investors in Coadjute include Lloyds, Nationwide, NatWest and Rightmove.
What about mortgage brokers?
Most industry participants accept that mortgage broking will change but probably not disappear. AI is likely to take over routine, standard cases, while more complex situations — for example freelancers, people with multiple income sources or first-time buyers — will still need human expertise. David Morris, head of residential lending at Santander UK, argues that success depends on responsible deployment of technology: combining human empathy and professional judgement with the speed and cost-efficiency AI provides.
Conclusion
AI is gradually embedded into UK mortgage processing and advisory services with the aim of accelerating processes and reducing costs. However, data quality, regulatory transparency and the handling of complex personal circumstances mean human advisers will remain important for complicated cases. The extent of digital transformation in the coming years will depend on how firms and regulators work together.
