Independent visual research / Public editionReviewed 04 Sep 2026
Motion fieldnotesOpenAI film study
About this study

Screen tutorial / 02:06 / Published 2026-04-24

Workspace agents in ChatGPT: Third-party risk management agent

A vendor-risk agent is built from an instruction and file, tested, and shown producing a structured document.

Watch official film ↗5 screen examplesAll 100 films
Representative frame from Workspace agents in ChatGPT: Third-party risk management agent
OpenAI · representative sample 01:48.000 · Source moment ↗

Useful to borrow

Use a test-before-publish chapter and end on the resulting evidence table.

Do not infer

The edited demonstration is not a legal, compliance, or risk-assessment validation.

Look closer

7 shot groups.

Read across from the observed content to its treatment and possible adaptation. Enlarge the frame for detail.

Approximate study intervals; horizontal scrolling reveals all columns on small screens.

Moment / sourceReference frameWhat is on screenFraming / editUseful adaptation
Shot group 01
≈ 00:00–00:26
App / browser screenWatch from here ↗
Sample 00:16.000
Risk-agent request

A setup prompt names vendor review and its sources.

Whole window then input crop.Make scope and sources visible.
Shot group 02
≈ 00:26–00:42
App / browser screenWatch from here ↗
Sample 00:28.000
Supporting file

A file picker adds reference material before build.

Native selection dialog.Show the evidence supplied to the task.
Shot group 03
≈ 00:42–01:10
App / browser screenWatch from here ↗
Sample 01:00.000
Build configuration

The builder reveals a plan and a split configuration surface.

Conversation-to-builder transition.Expose the resulting controls.
Shot group 04
≈ 01:10–01:14
Title / identityWatch from here ↗
Sample 01:12.000
Test chapter

A card explicitly introduces testing before publication.

Short visual reset.Make verification part of the story.
Shot group 05
≈ 01:14–01:46
App / browser screenWatch from here ↗
Sample 01:32.000
Test run

A preview conversation shows the request and task activity.

Closer preview panel.Distinguish a test from a live deployment.
Shot group 06
≈ 01:46–01:54
App / browser screenWatch from here ↗
Sample 01:48.000
Risk document

A document opens with a risk matrix and findings.

Result shown in its editor.Present the evidence artifact.
Shot group 07
≈ 01:54–02:06
Title / identityWatch from here ↗
Sample 02:00.000
Closing statements

Task categories and availability precede identity.

Grouped title cards.Keep claims bounded to what was demonstrated.

Reference frame

Enlarged reference frame

Watch this moment on the official channel ↗