Financial institutions are deploying agentic AI at scale without the governance infrastructure, authentication frameworks, or liability models to manage it safely. Yet governance and regulatory uncertainty represent greater barriers to scaling agent deployment than technology, talent, or cost. This reveals a critical insight: agentic AI adoption is not constrained by engineering—it is constrained by coordination and clarity.
Institutions face two simultaneous challenges. They are scaling internal agent deployments for fraud detection and AML while managing consumer-initiated agents they cannot reliably authenticate. Most lack formal governance policies, cannot reliably distinguish agents from humans, and cannot produce adequate authorization evidence if disputes arise. This evidence gap creates significant compliance and litigation risk.
This report examines why governance gaps represent the primary constraint on agentic AI scaling, what infrastructure institutions must build to manage agent deployments safely, and what regulatory exposure institutions face if they deploy without frameworks in place. Based on a survey of 155 fraud and AML executives across five markets, the research reveals concrete gaps institutions must close now before regulators mandate it later.
Clients of Datos Insights’ Fraud & AML practice can access this report.
About the Author
Other Authors
