EY Urges Decision-Governance Overhaul for Agentic AI Era

Agentic AI is shifting from an advisory tool to an operational actor capable of executing business processes and decisions with limited human intervention. That transition raises systemic risks because errors can propagate at machine speed. Traditional data governance, focused on information quality, access and compliance, may no longer be sufficient to establish accountability or maintain trust when automated systems act across an enterprise.

At the FinTechOn forum, EY Executive Vice President Tseng Yun said companies should expand oversight beyond data to encompass AI systems and the decisions they produce. She called for a multi-layer governance framework covering data, AI and decision-making, supported by clearly assigned responsibilities. Human supervision and circuit-breaker controls should remain in place so organizations can halt agents when they behave unexpectedly or operate beyond authorized boundaries.