Jessica Rusu's speech reveals the FCA's real 2028 timeline: AI agents will supervise approximately 60,000 anti-money laundering firms and 9,000 wholesale-market transaction-reporting firms using automated compliance tools. This is not hypothetical. For mid-market legal, insurance, financial services and accountancy firms already regulated by the FCA, this means your compliance posture must be readable, auditable and consistent enough for machines to assess it—not just humans. The regulator is not asking whether firms are ready. It is building the infrastructure on the assumption you will be.
This speech marks a watershed moment in how UK financial regulation works. We have moved from frontier AI (ChatGPT, Claude, Copilot) to agentic AI—systems that monitor, interpret and act on compliance data continuously. The FCA's evolution into an 'agentic regulator' mirrors what insurance and legal firms are attempting with tools like Harvey for contract review or Luminance for due diligence. But there is a crucial difference: the FCA's agents will be assessing you. The power asymmetry is stark. Firms that treat AI as a cost-saving tool will discover too late that the regulator's AI is designed to catch exactly the inconsistencies and workarounds that cost-cutting creates.
Trovix's view is uncompromising: most firms are preparing the wrong thing. They are buying general-purpose AI assistants or narrow compliance scanners and hoping these will somehow satisfy a regulator building purpose-built agentic systems. That approach will fail. The FCA's agents will not understand that your firm 'tried hard'—they will flag that your transaction-reporting data does not match your AML risk profiles, or that your client data was updated inconsistently, or that your controls produced contradictory outputs on Tuesdays. What matters now is not deploying AI to do work faster. It is implementing AI governance that makes your existing compliance controls auditable by machines. That requires Trovix Audit—a dashboard that gives you real-time visibility into how your firm's AI systems and human processes actually behave, so you can fix the gaps before the regulator's agents find them. It also demands Trovix Watch, because the FCA's supervisory rules for agentic regulation are still being written, and every change matters to how your controls will be assessed.
If you run a mid-market firm regulated by the FCA, your practical task is clear: audit your compliance data for consistency, logic and auditability right now. Do not wait for 2028. Do not assume your current tools are enough. The FCA Consumer Duty (PS22/9) already requires you to act in the firm's genuine interest; the PRA's operational resilience framework (SS1/23) demands you understand systemic dependencies; soon agentic supervision will require you to prove both through data and process patterns that machines can read. Start with your AML and transaction-reporting systems. Map which data points flow into compliance decisions. Identify where humans override rules and why. Implement governance that makes your reasoning transparent. This is not optional. It is the price of being supervised by AI.
Source: Financial Conduct Authority