Law360, London (October 19, 2026, 1:13 PM BST) -- The Financial Conduct Authority said Monday that it plans to become an "agentic regulator," using artificial intelligence to automate more of its supervisory work, analyze the activities of regulated companies and intervene more quickly when it identifies potential problems.


The FCA said this approach will help it respond to a financial sector increasingly reliant on rapidly advancing technology to provide services, make decisions and interact with customers.

The regulator said AI could help it examine firms' conduct and operations more efficiently, allowing the technology to "support rather than replace people's judgment."

But Rusu said becoming an agentic regulator would involve moving beyond using AI simply to assist employees to systems that can undertake more complex tasks and processes with greater autonomy.

"Agents can search across products, manage payments and investments, and transact directly with other agents, creating a new layer of autonomous technology between institutions and their customers," she said.

Rusu told a Treasury committee earlier in October that AI was particularly effective at analyzing and summarizing the "unstructured text" companies provide to the regulator in emails, documents and interviews.

She added that "becoming an agentic regulator is part of our overall ambition of being a smarter regulator." 

The FCA has also highlighted the importance of understanding how financial firms themselves are deploying AI. Its guidance draws attention to the need for firms to assess the risks of their systems and ensure proportionate oversight to protect consumers.

Data jointly published by the Bank of England and the FCA in 2024 states that adoption across U.K. financial institutions sits around 75% to 86%, with usage reaching up to 95% in the insurance sector.

The regulator is examining how existing rules should apply as responsibility for decisions becomes more dispersed among financial firms, technology providers and AI developers.

The FCA said it is building AI tools that will be operational in 2028 to oversee anti-money laundering measures in 60,000 financial services firms. It added that it will also deploy automation and agentic tools to supervise the reporting of approximately 9,000 small wholesale markets firms. The regulator said that this will allow supervisors to prioritize more complex activities.

But the regulator also warned that poorly designed AI systems could produce unfair outcomes, for example by excluding vulnerable consumers from services or making decisions influenced by bias. It said conducting live tests with financial firms had provided insights into both effective and flawed uses of AI, which it plans to use to inform its regulatory approach.

The FCA stressed that it is not proposing a separate set of AI-specific financial rules. Its approach remains focused on outcomes and is intended to be technology-neutral, meaning it aims to regulate what financial firms do and the outcomes for customers, rather than favoring or banning particular technology.

The FCA's long-term review, led by the regulator's executive director Sheldon Mills, is intended to help it understand how AI could transform retail financial services by 2030 and beyond. 

The review is examining how AI could affect consumers, competition and the relationship between customers and financial firms, as well as whether existing regulatory arrangements will remain effective as technology advances.

The FCA said it is investing £31.4 million ($41.8 million) in its core AI program between 2024 and 2027. It added that it expects investment in this area to increase further as it builds its AI capabilities.

Anca Thomson, corporate and commercial partner at Excello Law, said that while the FCA's good and poor practice examples "show where supervisors are looking," the more significant shift might be on the FCA's own side.

"If AI agents are to oversee some 9,000 transaction reporting firms, and tools are built for the 60,000 anti-money laundering firms joining its remit in 2028, the distance between a data error and a supervisory query is likely to shrink," Thomson said.

"Firms will reasonably expect the regulator to be as open about how its own AI is governed and tested as it asks them to be," she added.

Sophie Sheldon, AI lead for the financial institutions sector at Simmons & Simmons LLP, said there are concerns "around over-reliance on AI," particularly where firms become dependent on a single model, platform or provider. She added that these are "not an easy challenge to solve."

"Concentration risk, vendor lock-in and the absence of credible exit or substitution plans are likely to be key areas of regulatory focus as AI becomes more embedded in important business services," she said.

"The more difficult question, which isn't yet answered, is how the FCA expects firms to manage these risks in a market that is inevitably concentrated around a small number of major AI providers, where genuine substitutability may be limited," Sheldon added.

--Editing by Hazel Vidler. 

Update: This article has been updated to include detail about the FCA's AI investment program.

For a reprint of this article, please contact reprints@law360.com.