What happened
Published 26 May 2026 by the World Economic Forum in collaboration with Capgemini, this 40+ page playbook introduces the Agent Capability and Authorization Profile (ACAP) — a deployment-level governance and authorization instrument designed to make delegated agent decisions auditable, enforceable and accountable across the full lifecycle. The core argument is that organisations deploying AI agents face a structural gap: unlike human employees, autonomous agents lack implicit operational constraints, legal accountability, and reputational skin in the game, so governance must supply those boundaries explicitly and programmatically. The ACAP framework encompasses delegation policy, system design, and operational oversight in a single instrument — covering the degree of agent autonomy, the authority to execute and communicate, and the conditions under which human override is triggered. The report observes that many agents in a portfolio share the same foundational model, creating 'systemic vulnerabilities across the entire agent estate' if governance is not enforced at the instance level, and provides a structured path from individual pilots to governed portfolios at scale.
Why it matters
As agentic AI moves from experiment to enterprise-wide deployment — with agents handling pricing, supplier co-ordination, and hiring decisions in live operational settings — boards and CISOs need a concrete governance instrument, not just principles; ACAP gives them one they can map to existing risk and audit frameworks.
Action needed
Assign the CISO or Chief AI Officer to assess whether current agentic AI deployments can be profiled under the ACAP framework; identify the top five agents by consequence-of-failure and document their delegation policy and override conditions before next board cycle.