07/13/2026
Banks are starting to give AI agents identities, credentials and responsibilities. This is pivotal.
Reuters reported last week that major banks, including BNY Mellon, JPMorgan, Morgan Stanley and others, are expanding their use of AI agents beyond assistants to actively take part in work.
The question is fast becoming: How do you govern non-human workers?
Once an AI agent has access to systems, permissions to act and responsibility for completing work, key questions have to be answered:
• What is this agent allowed to do?
• Who approved those permissions?
• What happens if something goes wrong?
• Can every action be audited?
In Ovationcxm's recent survey of 522 U.S. finance leaders, we found the same pattern.
Organizations that want to use AI agents to complete banking tasks don't want unlimited AI connectivity.
They want connectivity with secure and compliant ex*****on... authentication, governance, confirmation, documentation and control.
Read all of our research into agentic banking and commerce ➡ https://www.ovationcxm.com/agentic-banking-report