FCA rethinks regulation for the AI age, eyes agentic watchdog

TL;DR:

• FCA chief Nikhil Rathi says the regulator is “re-thinking what it means to be an effective regulator” as AI moves faster than legislation can.
• The FCA is exploring agentic AI as a “first responder” to monitor wholesale markets across a billion rows of data a day.
• It points to tokenisation and agentic finance as the next scaling frontiers, backed by its Supercharged Sandbox and AI Lab.

The UK’s financial regulator is reframing its own role around AI. In a techUK speech, FCA chief executive Nikhil Rathi argued that with more than 80% of financial firms already adopting AI, the question is no longer whether to regulate for it but how to “preserve trust, competition, and resilience when technology is moving markets dramatically faster than the frameworks governing them”.

From rule-making to stewardship

Rathi’s central message was that “legislation will never keep up”, so a growing part of the FCA’s job is stewardship rather than detailed rules — and acting before laws catch up, as it did with Buy Now Pay Later. He flagged two scaling frontiers: agentic systems that “coordinate and transact” rather than just summarise, and tokenisation, citing the newly approved Baillie Gifford and BNY Mellon natively tokenised authorised fund. The regulator expects to use system-wide and competition powers “more frequently — not as exceptional interventions, but as a regular part of our toolkit”, insisting its role is “not to protect incumbents”.

The FCA is also turning AI on itself, exploring agentic AI as a “first responder” to spot market abuse faster across roughly a billion rows of data a day. On resilience, Rathi warned of mounting concentration risk across cloud, model and data providers — noting 98% of operational incidents reported last year were tech or cyber-related, and £1.3bn lost to UK payment fraud — making the Critical Third Parties regime “more important than ever”.

The stance is deliberately pro-innovation, echoing other UK regulators opening up, such as the MHRA’s London AI health sandbox, and it lands the same day a UK building society adopted agentic AI for underwriting — exactly the automated decisions Rathi wants kept under clear human accountability.

Looking forward

More is coming: the Mills Review on AI in retail finance lands within weeks, followed by guidance on good and poor AI practice later in 2026. For UK financial firms, the direction is a lighter-touch, collaboration-heavy regime that rewards responsible experimentation — a notable contrast with the more prescriptive EU AI Act, and a bet that “rebalancing risk” can attract AI investment without sacrificing market integrity.