Day-2 debrief | 85% say every AI agent has an owner. Only 42% say accountability is clear

85% of IT professionals said every AI agent had a named owner. Only 42% said accountability was clear (Ivanti, 2026). Can your organization prove who controls its production agents?

In this AI Day-2 Debrief, Matt Locknane and Rajeev Khanolkar examine the gap between assigning an AI agent owner and giving that person the visibility, access controls, and authority to act.

An agent can keep running after its sponsor changes teams. Delegated access can extend across other agents and tools. Older agent versions can retain active credentials long after their replacements arrive. These are practical AI governance and identity security problems that enterprises need to manage throughout the agent lifecycle.

In this clip:

• The four tests of agent ownership: knowing entitlements, reviewing activity, revoking access, and receiving alerts
• Why non-human identities can outlive their human sponsors
• How agent-to-agent delegation complicates accountability
• Why identity needs to follow the call chain across agents and tools
• Agent activity logs, regular reviews, and access monitoring
• Managing AI agent onboarding, operation, and retirement