AI in UK Mortgages & Housing Finance

Why this briefing

In June 2026 the FCA published two documents that belong side by side. Its economics function compared a conventional survival model with a gradient-boosted model on around 3.1 million first-time-buyer households and found the machine-learning model "attains a materially higher AUC (0.751 vs 0.669)" — "both models are equally accurate on average, but the gradient-boosted model is better at telling higher-risk households apart".

In the same month, its consultation CP26/18 described a borrower who had separated from a partner after "prolonged economic abuse", in stable employment and able to afford the repayments, "automatically declined by the firm because of their impaired credit history from this period". The better model and the automatic decline are the same technology. What separates them is whether a person can still see, and change, what happens to the individual.

Underneath both sits a number about to be produced at scale by models. Current loan-to-value drives arrears risk, and from 1 January 2027 the PRA's Basel 3.1 rules require a fresh valuation whenever a lender estimates a property has fallen more than 10 per cent, and at least every five years. The industry has told the PRA automated valuation models are "routinely used across the industry".

The human moment

The automatic decline — and, for the borrower already in the home, the forbearance decision that rests on a valuation they may never see. Since 4 November 2024 the FCA's rules have required support for borrowers in difficulty to be tailored to the individual. Since 5 February 2026 a solely automated decline has been lawful, with safeguards.

What's inside

Sixteen pages, written for an executive who knows the sector well and AI not at all:

• A note on the numbers — what is included, what was excluded, and why
• A £1.76 trillion market, growing again, with arrears falling
• The pressure point — fewer possession cases, each one taking longer
• The FCA's model comparison and its automatic-decline case, read together
• Four risks specific to mortgages, none of them about the technology working
• Where AI actually is — and what the largest lenders do not say
• The threat side — the title, the money at completion, and AI as the lock
• The UK legal position, and the Basel 3.1 valuation change nobody has framed
• Beyond the UK, stated as a comparator rather than an obligation
• Seven things that can be done now, none needing new law
• Three levels defined by where accountability sits, and three horizons
• An eight-workstream way in

The discipline behind it

Every figure carries a source. The FCA, the PRA, the Financial Ombudsman, the Ministry of Justice, HM Land Registry and UK Finance's own data carry the argument; where a lender appears, it appears from its own published record.

The most quotable claims in the sector are excluded on the record. Every statistic about AI-generated payslips and synthetic mortgage applicants traces to US trade press or vendor blogs, not a UK primary source. Vendor accuracy claims for automated valuation are excluded, and so is a widely quoted share of sales through brokers that traces only to a lender executive's blog.

Who it's for

A lender or intermediary executive who has just picked up an AI programme, a head of credit risk or collections, or a board member being asked to approve automation in a lending or forbearance decision. It assumes mortgages are well understood and AI is not.