FCA urges brokers to use AI - while keeping the human touch
The Financial Conduct Authority wants mortgage brokers to step up their use of AI and data tools to improve speed, accuracy, and access - without removing people from the process.
As part of building the "mortgage market of the future," the FCA highlighted four areas of focus. One of them - innovation and disclosure - pushes for "data and technology, such as AI, to help brokers give better and faster advice while keeping a human touch." Another calls for more flexible mortgage products to support first-time buyers and underserved groups.
David Geale, executive director for payments and digital finance at the FCA, said the regulator will "use insight from consumers and industry to drive further reforms and rebalance risk - helping to widen access to affordable mortgages to meet the needs of consumers today."
What this means for brokers and lenders
- Use AI for the heavy lifting: pre-qualification, policy matching, affordability checks, document parsing, and case summarization. Free humans for complex judgement and customer care.
- Keep humans in the loop: mandate adviser sign-off, escalation paths for vulnerable customers, and second-line review on edge cases.
- Tighten governance: data lineage, explainability, bias testing, audit trails, and periodic model validation. Align controls to Consumer Duty and mortgage conduct rules.
- Be explicit on disclosures: explain where AI is used, what it influences, and how customers can request a human review.
- Choose tech that plays well with your stack: CRM, sourcing systems, affordability calculators, and lender APIs. Prioritize privacy, security, and portability.
- Measure what matters: time-to-advice, case accuracy, offer rates, fall-out rates, complaint themes, and cost per completion.
Customers want speed - and a broker they trust
The finance sector is already using AI at scale, but the balance still matters. Research from Santander shows 52% of borrowers prefer a broker over AI for home buying, and 65% say they can't complete the process without one. Around 63% rely on real-world recommendations, while just 5% find their broker using AI. Most telling: 90% seek emotional support, and 82% say brokers reduce stress when things get complicated.
As Graham Sellar, head of intermediary channels at Santander UK, puts it: "AI provides support at the initial stages of deal research and exploration - but when it comes to making concrete life decisions, nothing can match the trust that is built through a broker."
Government push for smarter home-buying data
The government is backing a coalition, including the Centre for Finance, Innovation and Technology, to modernize the home-buying process and cut the cost of failed transactions. A smart data service and a potential digital platform are on the table to address collapsed chains, spiralling fees, and gazumping.
Practical 90-day plan
- Days 0-30: Map the end-to-end journey. Flag manual tasks, repeat data entry, and high-friction steps. Shortlist AI use cases with clear risk owners.
- Days 31-60: Pilot two use cases (e.g., document extraction and policy matching). Set guardrails: explainability, bias checks, adviser sign-off, and customer disclosures.
- Days 61-90: Integrate with core systems. Add monitoring dashboards. Train advisers on new workflows. Update policies and record-keeping to reflect AI involvement.
Compliance notes
- Document how AI influences advice and affordability assessments. Keep auditable records of prompts, outputs, and overrides.
- Run fairness tests across demographics. Calibrate models to reduce false positives/negatives that could impair access.
- Establish an appeals path and human override for advice that materially affects customer outcomes.
Resources
The signal is clear: automate the grind, protect the judgement. Brokers who pair AI with clear disclosures, measurable outcomes, and human-led advice will set the standard the FCA wants - and the market will reward it.
Your membership also unlocks:
