UK's FCA plans to become 'agentic regulator' using AI to automate financial supervision

The UK Financial Conduct Authority is moving to become an agentic regulator by integrating artificial intelligence to automate supervision, enforcement, and authorization workflows. While AI tools have reduced simple casework times from four hours to six minutes, CEO Nikhil Rathi resists establishing new AI regulations, relying instead on existing powers despite 20% of adults using AI for finance. Meanwhile, global watchdogs like the FSB and European Central Bank are tightening oversight on autonomous AI agents.

FCA agentic AI supervision

• ▪The Financial Conduct Authority completed a 12-week pilot project with Palantir testing artificial intelligence tools to integrate datasets and identify financial crime threats.
• ▪The Financial Conduct Authority's chief data officer Jessica Rusu stated the regulator aims to automate regulatory activities wherever possible and intervene at scale using artificial intelligence.
• ▪The Financial Conduct Authority plans to become an agentic regulator by using artificial intelligence to automate supervisory, authorization, and enforcement activities across financial services.
• ▪Jessica Rusu will deliver a speech outlining the Financial Conduct Authority's response to recommendations from the July 2026 Mills Review on artificial intelligence in finance.

AI automation efficiency gains

• ▪Jessica Rusu told UK lawmakers that artificial intelligence is particularly effective at analyzing and summarizing unstructured text provided in corporate filings and interviews.
• ▪The Financial Conduct Authority concluded artificial intelligence automation reduced simple casework processing times from up to four hours down to six minutes on average.

Regulatory perimeter gaps

• ▪Financial Conduct Authority chief executive Nikhil Rathi expressed skepticism about imposing new artificial intelligence rules, preferring to fully utilize existing regulatory powers first.
• ▪Financial Conduct Authority chief executive Nikhil Rathi stated to Parliament that artificial intelligence models providing general information without individualized transaction recommendations fall outside the existing regulated perimeter.
• ▪The July 2026 Mills Review urged UK authorities to evaluate whether large language models like ChatGPT, Claude, and Gemini should be subjected to financial regulations.

Consumer protection concerns

• ▪UK wealth managers expressed concern that consumers are seeking financial advice from unregulated artificial intelligence models without recourse to legal compensation if advice causes harm.
• ▪Research commissioned by the Mills Review found one-fifth of United Kingdom adults were open to using artificial intelligence models for personal financial decisions like borrowing or savings.

Global banking AI oversight

• ▪Banking regulators in the United States and Europe are questioning financial institutions regarding their usage of autonomous artificial intelligence agents and cyber defense systems.
• ▪The European Central Bank instructed Eurozone lenders to submit plans countering artificial intelligence cyber threats and map artificial intelligence usages across third-party suppliers by late October 2026.

FSB agentic AI guidance

• ▪The Financial Stability Board removed recommendations treating artificial intelligence as synthetic employees, warning that artificial intelligence agents do not reason like humans and may interpret goals unexpectedly.
• ▪The Financial Stability Board published final guidance on October 15, 2026, requesting increased human oversight and risk-based intervention triggers for autonomous artificial intelligence agents in banking.