How to version brand guidelines for AI workflows
Keep the current brand, saved task inputs and applied edits distinguishable. Use version references and a readable change record to diagnose stale guidance and reverse mistakes.
TL;DR
Save a distinct brand version when guidance changes, and record the version used by each task. Preserve the task's guidance and files where you need a lasting history. Inspect proposed edits before applying them, and write a short note explaining the change. Keep temporary campaign instructions with the task. Reverse a mistake through a new recorded change.
Keep today’s guidance separate from yesterday’s work
An old image may have followed then-current guidance, used a stale copy or ignored the right rule. Connect these records to tell the cases apart:
| Record | Question it answers |
|---|---|
| Current brand | What should a new task use now? |
| Saved task inputs | Which guidance and files did this output use? |
| Applied change | What changed between those versions, and why? |
Preserve completed work's original inputs as the current brand changes.
Save a coherent version of the brand
- Keep the package together: guidance, identity values and asset references. Recording only the voice document leaves logo and layout decisions unexplained.
- Save changes separately: follow Semantic Versioning's rule that a released version remains unchanged.
- Write a dated change note: say what was corrected, added or retired. Software can retain exact references without showing them to readers.
- Archive files when needed: a saved version does not prevent a temporary download link expiring.
Attach the version to the task when it runs
Rainbrand maintains brand context for agents and products. A Brand Skill saves the complete guidance and asset references; a Brief selects guidance and files for a task.
Store the version identified by the Brief response with the task and output. A practical record is:
Output file or content record:
Task:
Brand:
Guidance used for this task:
Where to find that saved guidance:
Campaign instructions supplied:
Guidance and selected files, or archive location:
Producing tool or model:
Review findings and revisions:
This is a workflow record, not an API response schema. Save it during the work; a name and timestamp may not establish which context reached the producing tool.
Prepare and inspect the next change
- Prepare a Brand Edit. It proposes changes without changing active guidance.
- Compare the proposed files before and after, then refine, apply or discard through the authorised workflow.
- If another edit changed the brand first, the stale proposal returns a conflict. Prepare it again from the current version.
Record direct Library additions and removals too. Removing a file stops future selection while previously published Brand Skills keep working.
Explain what the new version changes
A change note tells colleagues what to do. Follow Keep a Changelog: dated, readable entries, latest first, with related changes grouped.
What changed:
Why and which source established it:
When new work should use it:
Active work that needs revising:
Files or exports to replace:
Keep current guidance readable without reconstructing it from the changelog. After applying an edit, request a fresh Brief; completed Briefs stay unchanged. Use the cross-tool checklist for retained exports and instructions.
Use the record to diagnose a wrong output
Compare the saved inputs with the output:
| Finding | Action |
|---|---|
| Older guidance was intentional | Preserve it with the historical work. |
| An old copy reached work needing today's brand | Retrieve fresh context and fix the copy or reuse decision. |
| The intended version contained a wrong rule | Correct the brand and find other affected tasks. |
| A correct rule was ignored or its file omitted | Correct the output or handoff, not the brand rule. |
Keep Review findings with the checked output and guidance. Review supplies findings; creation tools make corrections. A later check must not overwrite the original task's history.
Reverse a mistake without erasing its history
Restore a rule through a new edit, preserving the version that introduced the mistake. Like Git revert, this records the reversal rather than erasing it. Inspect, apply and retrieve fresh context for affected work.
Record why and which outputs need revising. For a permanent archive, export the Brand Skill and retain task guidance and files too: deleting a saved version can remove a Brief's version link.
Keep temporary instructions with the work
Keep campaign headlines, offers, dates and usage restrictions with the task and its Brief request, unless they are intended as lasting brand changes.
The runtime guide explains comparing current and saved versions before reuse. Start by finding one recent output's task, original guidance and supplied files together; retain that history as new work uses the updated brand.
More in Brand guidelines
All 7 Brand guidelines articles6 min read
Brand guidelines vs brand context: what does an AI agent need?
Brand guidelines explain the whole brand to people. Brand context is the part of it an agent receives for one task, in the form that task uses. Here’s how one brand becomes different contexts for different tasks, why more context isn’t better, and how to keep it current.
4 min read
How to turn a brand PDF into instructions AI can use
Turn a brand PDF into explicit rules, exact values and usable asset files. Inventory the guide, resolve conflicting instructions, and keep the result current across agents.
4 min read
How to turn your website into brand guidelines AI can use
Use your website as evidence for brand guidelines. Learn what to extract, how to resolve gaps and conflicting choices, and how to keep agents current.
