Schedule
Brand Guard has no schedule. It does not scan your content in the background or post on a timer. It runs when another agent asks it to check a draft, and when you ask for an audit or topic suggestions.
When It Runs
| Situation | How it runs |
|---|---|
| A publishing agent approves or publishes a draft, with brand checking set to Warn or Enforce | Automatic. The agent asks Brand Guard for a verdict before it goes on |
| The same situation with brand checking set to Off | It does not run |
| You start an audit with New Audit in the app | On request, in any mode |
| You run the suggest-topics or review-topics command in Slack | On request, in any mode |
Brand checking mode is set in the Operator Console. Warn is the shipped default, so out of the box every approval or publish is checked and the results are recorded, but nothing is held back. Turn on Enforce to hold failing drafts, or Off to skip the check.
What Happens in a Check
The fixed rules run first, since they need no judgment and settle the obvious cases. Then the draft is read against the brand's voice and scored. If a fixed rule found something serious, or the first read is borderline, a more careful pass can run before the decision. The calling agent waits for the verdict before it goes on.
Re-Checking After an Edit
Brand Guard does not watch for edits. After you change a draft, the producing agent runs the check again the next time you approve or publish it, or you can paste the new text into New Audit to see the result first.
Topic Suggestions
Brand Guard proposes topics only when you run the Slack command. Ideas wait on the card, and in your topic queue, until someone Greenlights or Dismisses them.