The FCA's AI Roadmap: Why Mortgage Brokers Can No Longer Afford to Ignore Unregulated Advice
8 min read
Introduction
The mortgage landscape is experiencing a subtle but profound structural shift in how consumers interact with financial guidance. For years, the traditional broker model relied on being the sole gatekeeper of product information, affordability criteria, and market execution.
However, the rapid consumer adoption of general-purpose AI has created a parallel, unregulated advisory market right under the noses of regulated firms. The regulatory response has finally arrived, and it should serve as a wake-up call for every mortgage operator in the UK.
1. The Regulatory Context: Understanding the Mills Review
Published by the Financial Conduct Authority (FCA) in July 2026, the Mills Review is a carefully balanced document. It acknowledges risks, recommends safeguards, and is written in the measured language of a senior regulator who spent months consulting with industry. The FCA has thought hard about the future of AI in retail financial services.
Mortgage brokers, however, should read it with a sense of urgency.
The reviewâs central finding is that the structural shift in financial services is moving rapidly from human-led, episodic activity toward services that are AI-enabled, continuous, and delegated. That shift is not a future projectionâit is already well under way:
Widespread Industry Adoption: A joint survey by the Bank of England and the FCA found that 75% of UK financial services firms are already utilizing AI within their operations, a sharp increase from 58% in 2022.
The Shift to Autonomy: The Mills Review's consumer research revealed that one in five UK adults (20%) is already completely open to letting AI make financial decisions for them outright.
High-Risk Verticals: Demand for autonomous AI decision-making is currently strongest in debt advice, pensions, and investmentsâplacing the mortgage market directly in the crosshairs of this behavioral shift.
2. The Recourse Mirage: The 26% Accountability Gap
The most alarming statistic lands heavily on page five of the executive summary: around 26% of UK adults trust general-purpose tools such as ChatGPT, Claude, or Gemini for financial advice.
Crucially, this group exhibits a dangerous blind spot: they have limited to no awareness that formal routes to regulatory recourse will not apply to these interactions.
The Consumer Recourse Void: One in four of your prospective clients may be taking structural mortgage guidance from an unregulated, general-purpose large language model. When these tools output misinformationâand they frequently doâthe consumer enters a complete regulatory void. There is no Financial Ombudsman Service (FOS), no professional indemnity insurance protection, and no formal FCA complaint process. There is simply zero recourse.
Denni Tyson, founder of DT Financial, has already witnessed the real-world fallout of this trend. "The level of misinformation that is out there is quite frightening now for people," she notes. "The concern is the fact that people maybe start relying on it and it makes a mistakeâwhich the technology openly admits it can doâand there's just no accountability. You cannot blame a computer when your mortgage application is declined or you lock into the wrong product."
3. The Regulatory Perimeter Problem and The Consumer Duty Trap
The Mills Review highlights a severe structural challenge: the regulatory perimeter problem. General-purpose AI tools are actively shaping high-stakes financial decisions and altering market competition without clear FCA oversight. Because whether their activity falls inside or outside the regulatory boundary is currently unresolved, the review recommends securing and adapting that perimeter as its first priority.
However, regulatory recommendations take years to turn into enforceable law. In the meantime, consumers are using ChatGPT for active mortgage research today, creating a massive compliance trap for human advisers.
Under the Consumer Duty obligations introduced in July 2023, the responsibility for delivering and demonstrating good outcomes lands squarely on the shoulders of regulated firms.
If a client approaches a brokerage having made a critical financial assumption based on AI-generated hallucinations, the burden of unpicking that error falls entirely on the human adviser. The AI tool bears no liability; the regulated broker inherits it all.
Rhys Edwards, mortgage consultant at Brooks Financial, emphasizes the baseline reality: "You can't fully rely on it just because an AI states that a lender might accept a specific complex income structure or that a certain rate is available online." Yet, a significant portion of the public has not yet reached that logical conclusion.
4. The Micro-Economics of the UK Advice Gap
While the risks are stark, the Mills Review emphasizes that the driver behind AI adoption is a massive, systemic failure in the traditional advice market. The "advice gap" in the UK is real, large, and remarkably expensive:
Exclusivity of Advice: Only 9% of UK consumers currently access traditional, human-led financial advice.
Stagnant Capital: An estimated £300 billion sits idle in low-interest accounts, losing value to inflation.
Financial Exclusion: Around 900,000 UK adults remain entirely unbanked.
The regulator's core argument is that AI is uniquely positioned to close these gaps by securely blending a consumer's open banking data with broader market knowledge to deliver instant, highly personalized recommendations.
On mortgage switching specifically, the review identifies historically low retention and switching rates as a prime inefficiency that automated infrastructure will inevitably solve.
5. The Autonomy Spectrum: Adapting vs. Expiring
Current consumer sentiment shows a lingering hesitation toward pure automation, but the trajectory is clear:
Pure Automation Trust: Only 12% of buyers state they would fully trust a standalone, completely automated AI mortgage platform.
Human Preference: 30% of consumers still actively prefer receiving advice directly from a human professional.
However, the Mills Reviewâs Autonomy Spectrum suggests these numbers will shift rapidly as AI systems become more deeply embedded into everyday digital banking ecosystems.
The consumer who today uses an LLM to compare basic products before contacting a broker will, by the review's own projections, increasingly delegate the entire journeyâcomparison, recommendation, and rate lockingâto verified autonomous systems. Human involvement will steadily contract from an open-ended conversation to a simple, digital sign-off.
Era of Acquisition | Primary Channel | Human Role | Operational Speed |
Traditional Model | Phone / In-Person | Full Intake & Product Sifting | Days to Weeks |
Hybrid Model | Web Forms / Portals | Manual Follow-up & Triage | Hours to Days |
The AI-Enabled Future | Autonomous Concierge | Strategic Sign-off & Complex Advice | Real-Time / Instant |
As Babek Ismayil, founder of OneDome, plainly states: "AI will not replace mortgage brokers. But it will replace firms that fail to adaptâfirms that continue to think about mortgages in isolation rather than the broader digital customer journey."
Conclusion: The Broker's New Real Estate
The long-term threat to the mortgage profession is not a sudden, overnight replacement of human advisers by technology. Rather, it is the gradual absorption of the routine, transactional parts of the roleâproduct comparison, basic affordability indicatives, and automated rate alerts.
As automated infrastructure perfectly optimizes the front-end customer journey, the space available for human value will shrink exclusively to complex underwriting cases, specialized lending structures, and high-empathy client management. Brokerages that fail to integrate seamless inbound infrastructure to capture and guide these clients early will find themselves entirely cut out of the conversation.
