The AI Input Zone Closed: The FCA Now Has the Industry's Evidence

UK financial services firms are operating under one of the most demanding accountability frameworks in the world. The FCA, the Financial Conduct Authority, the UK's main financial regulator, has been explicit: it does not intend to write new AI rules. What it wants instead is evidence of what firms are actually doing, so it can publish guidance that becomes the effective standard the industry is measured against. This week, two things happened simultaneously.

The FCA's AI Input Zone, the evidence-gathering exercise where regulated firms were asked to submit specific examples of their AI governance, closed Friday, 19 June 2026. And on 10 June, the FCA published its first-ever Emerging Technology Horizon Scan, a map identifying agentic AI, synthetic fraud, and programmable finance as the headline risks for financial services. Under SM&CR, the Senior Managers and Certification Regime, named individuals at your firm are personally accountable for each area of the business they oversee.

The FCA now has the industry's evidence on AI governance, and it has published its own map of what it is watching next. The gap between those two things is where accountability will be tested.

The AI Input Zone Closed on Friday 19th June: The FCA Now Has the Industry's Evidence

The FCA's AI Input Zone reopened in May 2026 and asked firms for something specific: not opinions on AI policy, but concrete examples. What has enabled safe and responsible AI deployments? What has blocked them? What should the regulator address in its forthcoming good and poor practice publication on AI? Responses were due Friday, 19 June 2026. The FCA has stated it will anonymise all submissions before publication. The good and poor practice publication is expected later in 2026 and will be the first substantive FCA output setting expectations for AI governance across regulated firms.

The firms affected are all FCA-regulated firms that have deployed AI in financial services: retail banks, insurers, consumer credit providers, investment managers, wealth managers, and payment firms. Any firm that chose not to respond has, in practical terms, handed the framing of good practice entirely to its peers. The examples that shape the publication will be drawn from firms that engaged. The firms that did not engage will be compared against a benchmark they had no hand in setting.

The AI governance implication is direct. Under SM&CR, a Head of AI or Head of Compliance whose firm's practices do not align with the published good practice examples has a documented gap. That gap does not live in a policy paper. It lives in their Statement of Responsibilities. The FCA has not written new rules for AI. It has done something more targeted: it has asked the industry to describe good practice, it will publish those descriptions, and it will then supervise against them. The standard is being set this week by the firms that responded. The supervision will follow.

They optimized for LLMs, not just Google.

FAQs. Comparison pages. Transparent pricing. LinkedIn presence. These aren't vanity plays. They're what gets you cited in ChatGPT, Gemini, and Claude when your buyers are researching, your investors are looking, and your future hires are deciding where to work.

Download the free AEO Playbook for Startups from HubSpot and get the exact checklist. Five minutes to read.

Unlock the playbook

The FCA's Emerging Technology Horizon Scan 2026: Agentic AI Is Now on the Map

On 10 June 2026, the FCA published its first-ever Emerging Technology Horizon Scan. It is not regulatory guidance. The FCA is explicit: the document does not set rules or make predictions. What it does is publish where the regulator's analytical attention is going, which is the next best thing to guidance for any firm trying to understand supervisory priorities. The three headline themes are Personalised Intelligence, where AI agents are reshaping how consumers interact with financial services; Synthetic (In)security, where deepfakes and AI-generated fraud are outpacing traditional firm controls; and Programmable Finance, where tokenisation, stablecoins, and distributed ledger technologies are moving from pilots to live environments.

The firms affected span the breadth of UK financial services. Retail banks and consumer credit firms are exposed through Personalised Intelligence, as AI-driven financial guidance and automated customer interaction become mainstream. Wealth managers and investment firms face the same. Firms in crypto and payments face the Programmable Finance chapter directly. Every firm faces the Synthetic (In)security risk: the FCA notes that synthetic media, including audio and video deepfakes, is becoming harder to distinguish from real content at a pace that outstrips existing fraud controls.

The AI governance implication sits in the Horizon Scan's treatment of agentic systems. The FCA states that agentic AI can execute thousands of transactions or decisions in seconds, and that if these systems are misaligned or manipulated, the velocity of harm outpaces traditional firm oversight, making detection and remediation difficult. This is a regulatory signal, not a rule. But regulatory signals from the FCA do not stay signals for long. The sequence is predictable: horizon scan, thematic review, Dear CEO letter, enforcement. Under SM&CR, every decision that touches a regulated activity needs a named senior manager who can explain the oversight framework for that system. For agentic systems operating at the speed the FCA is now describing, the question is not whether a human approved each individual transaction. The question is whether a named individual understood the parameters of the system, documented that understanding, and had a genuine control mechanism to stop harm once it was detected. Most firms using agentic AI cannot answer that question today.

Two documents published within the same week. The FCA now has the evidence firms submitted about their AI practices, and a published map of where it believes the biggest technology risks are emerging. What this week exposes is the mismatch between the speed at which AI is being deployed in UK financial services and the speed at which human accountability frameworks are being updated to match it.

Agentic AI systems executing thousands of decisions per second are being governed by accountability frameworks built in 2016 for humans making individual choices. A Head of Compliance or Head of AI who reads the Horizon Scan should not file it as background reading. They should map it directly against their firm's AI inventory: for each system the FCA has flagged as a horizon risk, is there a named senior manager with documented oversight? If the answer is no, the good and poor practice publication expected later this year will make that gap visible in ways a supervisory visit will not miss.

For each AI system your firm operates that functions autonomously, executing decisions without individual human approval at the point of action, can you name the senior manager accountable for it, and can you show the regulator exactly where that accountability is documented in their Statement of Responsibilities?

If this is a live problem in your organisation, reply to this email. BuildMoat works with UK financial services firms to build the evidential accountability you need to stay confident even when the regulator calls.

BuildMoat — buildmoat.co.uk

That's it for this week. If something landed, reply and tell me. If something was wrong, reply and tell me that too. I read every response.

Forward this to one person who should be reading it.

Find me on X and LinkedIn if you want to continue the conversation there.

For sponsorship enquiries, email [email protected]

Regulatory commentary is for informational purposes only and does not constitute legal, compliance, or financial advice. Consult a qualified professional for advice specific to your firm. BuildMoat product mentions are editorially assessed and disclosed where a conflict exists.

Until next week.