AI agents are gaining authority across enterprise systems. But are organizations prepared to monitor their actions and establish accountability when something goes wrong?

SecurityHQ, in partnership with Datadog, surveyed 204 cybersecurity leaders to understand how organizations are adopting, securing, and governing autonomous AI agents.

The findings reveal a growing disconnect between the rapid expansion of AI agents and the operational capabilities needed to govern them.

While 93% of respondents believe AI agents will access enterprise systems more frequently than humans within five years, only 49% say they could fully reconstruct every action an agent performed over the previous 30 days.

The joint research from SecurityHQ and Datadog also highlights a gap between governance policies and operational accountability:

  • 76% have formally defined accountability for AI-driven decisions.
  • 45% have experienced an AI- or automation-related incident where accountability was unclear.
  • Only 17% believe they could reconstruct every action in an agent-driven critical incident within minutes.

These findings raise important questions for security leaders. As autonomous systems gain greater access to infrastructure, applications, and sensitive data, how can organizations maintain visibility into their actions, investigate incidents, and ensure accountability?

Our new report with Datadog, The Accountability Gap: The State of AI Agent Security and Governance 2026, explores these challenges, including shadow agents, machine identities, human oversight, and the growing complexity of governing AI across enterprise environments.

Download the Full Report

Explore the findings and what they mean for the future of AI-agent security and governance.