Financial Services’ next AI risk is the workflow nobody can explain (www.techradar.com)

Financial services are transitioning from using AI for answering questions to implementing AI that takes decisive actions, as outlined in the Government’s Financial Services Adoption Plan. This shift is significant because it changes the landscape of accountability; firms must now understand not only how decisions are made, but also the subsequent actions taken by AI agents. The Financial Conduct Authority highlights the rising demand for explainability in AI workflows, where the complexity increases as these systems are set to autonomously coordinate and transact across various processes, such as fraud detection and compliance operations. As nearly half of UK financial services executives report using agentic AI, concerns emerge around governance and compliance—over a quarter have limited controls in place. This agent sprawl can lead to fragmented decision-making processes, where traceability suffers, making it difficult to pinpoint accountability when things go wrong. The call for enhanced oversight is clearer than ever; firms need comprehensive records of every interaction between AI agents and the systems they interface with to ensure transparency, as well as robust mechanisms for intervention to prevent potential harm. As AI becomes more integrated into core workflows, the challenge will be to maintain visibility and governance, enabling the benefits of AI while safeguarding accountability.