Five banks blocked third-party AI agents from operating inside customer accounts. They cite risks like fraud, unauthorized transactions, privacy, liability and regulatory accountability. AI agents can act autonomously and amplify mistakes or malicious activity. Blocking is a short-term fix but not a long-term strategy. Customer demand for fully autonomous financial agents remains limited today. Half of US and Canadian consumers do not trust AI with sensitive financial information. Forty percent have concerns about data security and confidentiality. Banks currently allow third-party agents read-only access via approved providers. Banks risk becoming commodity product providers if they abstain from agent interactions. Forrester recommends a graduated trust framework for agent access. Banks should authenticate customers and agents and verify agent provenance. Permissions should be explicit, revocable, time-limited and tied to defined actions. High-risk activities should initially require customer approval.

• 40% have concerns about data security and confidentiality