Skip to content
Blog
Azure AI FoundryAI radar

Foundry highlights human approval in long-running agents

The August documentation roundup includes a human-approval guide for long-running agents, marked as preview.

Prepared on Oct 28, 2026 · 1 min

Source documentation update:

What the official source reports

The August documentation roundup includes a human-approval guide for long-running agents, marked as preview.

What this means for your operation

Our view: a process lasting hours may need to wait for a decision, preserve context and resume. Workflow design should identify who approves, what they see and what happens if nobody responds. Human involvement helps only when the decision is understandable and has a defined consequence.

Next step for evaluation

Choose a consequential action, such as changing an order, and specify the minimum approval information. Test rejection, expiration and resumption before the happy path. Confirm provider preview restrictions and environment region. This news does not mean that this particular implementation is already incorporated into Zild Flow.

Sources and references

Retrospective edition 2026-W26. The week identifies this collection’s editorial organization; the text was prepared on the date shown above.

Bring this idea into your operations.

Map the process, connect systems and evaluate outcomes with Zild.

Talk to an expert