The Deutsche Bundesbank warned in its September 2026 report that the widespread adoption of AI agents in corporate treasury could lead to systemic risks. The use of homogeneous models and identical data sources may cause thousands of independent systems to make identical financial decisions simultaneously, creating collective herd behavior.
Such synchronized actions could result in payment gridlock, sudden liquidity demands, and infrastructure stress. The Bundesbank noted that while individual companies may achieve efficiency, the aggregate effect of autonomous agents executing transactions without human approval could amplify market movements and create procyclical risks.
Data from PYMNTS Intelligence shows that as of September 2025, 7% of US enterprise CFOs had deployed agentic AI in live finance workflows, with another 5% running pilots. A separate August 2025 report highlighted a readiness gap, with 75% of tech firms reporting familiarity with agentic AI, compared to 33% of goods firms and 38% of services firms.
To mitigate these risks, the Bundesbank recommends implementing clear mandates, reliable identity verification, and human intervention mechanisms. CFOs are advised to establish escalation triggers and checkpoints for large transactions, ensuring that automated systems have defined limits on their autonomy when market conditions become volatile.
