What do the FCA’s latest AI publications mean for your firm?

The FCA wants to support the safe and responsible adoption of AI, with firms maintaining clear accountability, effective governance and evidence of good customer outcomes. Its latest publications do not introduce new AI-specific rules but explain how existing requirements, including the Consumer Duty, SM&CR and operational resilience, apply as firms move beyond experiments and use AI more widely. 

AI

AI is taking a larger place on the FCA’s agenda as firms are exploring how to use it in customer interactions and communications, advice, analysis and day-to-day operations. 

In October 2026, the FCA explained its position in more detail through two communications. Its publication, AI in financial services: good practice and areas for improvement, gives practical examples of what the regulator is seeing in AI strategy, governance, risk management and consumer outcomes. 

In her speech, Building the future of finance in the age of AI, Jessica Rusu, the FCA’s Chief Data, Information and Intelligence Officer, described the regulator’s wider approach as AI becomes more embedded in financial services. She also looked ahead to more autonomous, agentic AI. 

Together, these publications give firms a clearer sense of what the FCA expects from responsible AI adoption. There is no new set of AI-specific rules, but existing regulatory expectations still apply as the technology and its uses change. 

What does good AI governance mean in practice?

The FCA’s communications return to three themes: consumer outcomes, accountability and operational resilience. 

Starting with the customer need, it’s important that firms are able to explain what problem they want AI to solve and why it is the right tool for the job. The FCA identifies adopting AI without first defining the consumer need, target market or foreseeable harms as an area for improvement. 

The Consumer Duty remains central too. Firms should test outcomes for customers with different characteristics and across different scenarios, check for bias and accuracy issues, and monitor systems once they are in use. Customers in vulnerable circumstances need particular consideration. 

For smaller firms, this needs to be proportionate. The FCA’s examples recognise that testing at scale is not always feasible. Smaller firms can, for instance, review sample outputs, add vulnerability markers to test cases and compare results across different scenarios. 

Who is responsible when firms use AI?

Using AI does not remove individual or organisational accountability. The FCA expects material AI uses to have clear owners, appropriate risk-based controls, effective oversight and arrangements for escalation. Firms therefore need to understand how AI is used across the business, not just which tools they have formally bought. 

An AI inventory is a useful place to start, recording the tools in use, their purpose, provider, owner and risk classification. Firms can then map material use cases to accountable senior managers and keep proportionate records of approvals, oversight and testing. Human oversight is an important safeguard but firms should not rely on it alone to manage AI risks – it should work alongside well-designed technology, testing and controls. Reviewers need the expertise and authority to challenge or override outputs. How much they intervene should depend on the risk of the use case. 

What about AI supplied by third parties?

The same responsibility applies when firms use third-party AI. Firms should understand their dependencies and concentration risks, test supplier assurances, and maintain appropriate fallback and exit arrangements. Outsourcing a tool, model or platform does not transfer responsibility for customer outcomes. 

These questions become more important as AI gains autonomy. The FCA is already considering “agentic finance”, where AI could research products, manage investments and payments, or execute transactions. It intends to work with firms, other regulators and AI developers on agent identity, delegated authority, accountability and execution controls. 

What should firms do now?

These publications do not introduce new requirements or an associated implementation deadline, but firms can use the examples to assess whether their current and planned AI arrangements remain appropriate.  

Practical steps could include: 

  • keeping an up-to-date inventory of AI tools and material use cases 
  • making ownership and senior-management accountability clear 
  • checking customer outcomes, potential bias and the needs of customers in vulnerable circumstances 
  • keeping proportionate records of testing, decisions and changes 
  • giving human reviewers the ability to challenge AI outputs meaningfully 
  • defining risk tiers and thresholds for escalation 
  • reviewing third-party dependencies, resilience and fallback arrangements. 

The FCA’s position is becoming more defined. Firms do not need elaborate governance just because a process uses AI. They do need to be able to show that their controls are proportionate to the risks and work in practice. 

Supporting sources

  1. Building the future of finance in the age of AI
  2. AI in financial services: good practice and areas for improvement

Frequently asked questions

Are there new FCA rules on AI for financial services firms?

No. The FCA says its examples neither create new requirements nor prescribe a single governance model. Firms must still meet existing requirements, including the Consumer Duty, SYSC, SM&CR and operational resilience requirements. 

Does AI governance need to look the same in small and large firms?

Not necessarily. The FCA stresses proportionality and recognises that smaller firms may not have the resources to test AI on the same scale as larger firms. Governance and testing should suit the firm’s size, the use case, the risks and the potential impact on customers. 

What are the first steps in reviewing AI governance and oversight?

Start by identifying current and planned AI use and agreeing who owns it. Then assess those use cases against the FCA’s examples, with particular attention to the Consumer Duty, governance, human oversight and operational resilience. 

Ready to strengthen your compliance?

Speak to our experts about your regulatory challenges.