Financial institutions face a growing struggle to tell human clients apart from automated software. A 2026 survey by behavioral biometrics firm BioCatch revealed that 80% of banking compliance and fraud professionals across 25 countries have already encountered attacks using agentic artificial intelligence.
The poll of 1,440 risk specialists found that 84% view AI agents as the top exploitable industry vulnerability for the coming year. Furthermore, 72% expect it to be very difficult to separate legitimate AI tasks from malicious actions. Major UK banks, including HSBC, Barclays, NatWest, Lloyds and Nationwide, have reportedly blocked AI tools from accessing client accounts due to fraud and liability worries.
Retail brokers take a different approach. MetaQuotes, Spotware and other major platform providers now let users link trading accounts to AI agents. Robinhood requires clients to open a separate agentic account and puts full financial responsibility on the user for any automated activity. BioCatch recently launched a detection capability named AiRE to spot behavioral anomalies and software entities before funds move.