- UK banks are entering a race to deploy Agentic AI, a type of AI capable of planning, decision-making, and acting independently for retail customers.
- The Financial Conduct Authority (FCA) warns this technology brings new risks to consumers, despite its potential to transform how people save, invest, and manage money.
- Unlike generative AI, which only creates text, code, or images, Agentic AI can automatically transfer idle funds to high-interest accounts or adjust portfolios during market volatility.
- NatWest, Lloyds, and Starling are working with the FCA to prepare tests of Agentic AI for customers, marking a shift from back-office AI to direct applications.
- The FCA expects consumer-facing Agentic AI applications to appear clearly from early 2026, but has no plans for new regulations yet.
- Existing rules like the Senior Managers Regime and Consumer Duty will be used to hold bank leaders accountable and prioritize customer interests.
- Gartner predicts 40% of financial firms will use AI agents by late 2026, though 40% of projects may be canceled before 2027 due to high costs and unclear business value.
- Lloyds is testing AI that helps customers automatically invest savings into tax-free ISAs with prior permission.
- Starling is developing personalized budgeting tools, setting up standing orders and predictive spending limits.
- The biggest risk lies in multiple AI agents interacting simultaneously, potentially reacting identically to market signals, which could increase the risk and speed of bank runs.
- Agentic AI remains weak with complex tasks, can hallucinate, and poses challenges regarding reliability and bank leadership’s understanding.
📌 Summary: UK banks are racing to deploy Agentic AI, which can plan and act independently for retail customers. Unlike generative AI, it can automate fund transfers and portfolio adjustments. Regulators warn of new consumer risks even as the technology promises to revolutionize financial management.
