Authority latency
The elapsed time between a workflow becoming ready for a consequential decision and an authorized person or system actually taking that decision or action.
Authority latency is the elapsed time between a workflow becoming ready for a consequential decision and an authorized person or system actually taking that decision or action.
It differs from ordinary process or technical latency. A workflow may already have the information it needs, yet still wait because the right approver is unavailable, ownership is unclear, a threshold has not been delegated, another function must sign off, or the escalation path is ambiguous.
The term becomes especially useful in agentic workflows. AI can reduce the time spent gathering, reconciling and routing information while leaving existing approval structures unchanged. In that case, authority latency can become the dominant remaining bottleneck even though the underlying workflow has become technically faster.
Authority latency should not automatically be minimized. High-consequence or irreversible decisions may deserve deliberate delay. The useful goal is to make the delay intentional, measurable and proportional to risk.
It relates to task-scoped authority, escalation, threshold policy and the operating model.
Read more in AI is removing the coordination tax. The next bottleneck is authority.
Related terms
Task-scoped authority
A temporary authorization envelope for one AI-agent task, limited by allowed action, resource, consequence, duration and recovery conditions rather than broad standing role permissions.
Escalation
Passing a case from an automated system to a person (or a higher level) because it cannot be resolved safely or well. Handled early, it is protection, not failure.
Threshold policy
The explicit rules that turn a model's probability or score into an action — reject, send to a human, or act autonomously — owned and reviewed separately from the model itself.
Operating model
How an organization actually runs: its workflows, decisions, roles, governance and handoffs. AI creates value when it changes the operating model, not when it is bolted onto the old one.