← Blog

Series

Agent approvals

Choose which agent actions can finish automatically and which need a human decision.

5 articles

Reading list

In this series

Browse every article →
  1. 01

    · Approvals & policy · 9 min

    Designing approvals operators actually read

    The failure mode every agent-approval system hits is rubber-stamp clickwrap. Design moves that delay it: signal selection, payload composition, deadlines, audit, and the hand-off problem.

  2. 02

    · Approvals & policy · 10 min

    Policy at action time: what the gate evaluates

    The five inputs and three checks a runtime policy gate needs to decide whether an agent action should run, with policy-language and approval-UX considerations.

  3. 03

    · Approvals & policy · 4 min

    Controls that hold when the agent stops agreeing

    An instruction lives inside the agent's context and competes with everything else there. A gate lives outside it, at a boundary the agent cannot cross without a decision it doesn't control.

  4. 04

    · Approvals & policy · 5 min

    Every content exception needs an owner before it becomes a cost

    A practical exception process for AI-assisted marketing work that is missing a source, owner, approval, budget, or safe path to publication.

  5. 05

    · Operations · 4 min

    Your AI program has a review queue, even if you call it approval

    Measure waiting time, reviewer effort, and rework in the queue between an AI-assisted draft and a published marketing asset.