Create·Advanced·30 min·Updated Sep 30, 2026

Auto-check campaign content against brand and regulatory rules

Build a Compliance Check agent that verifies brand and regulatory alignment before campaign content ships.

Download PDF

Microsoft 365

Works With

Prerequisites

Brand and regulatory guidelines document

Business Outcome

Brand and compliance issues caught before publication, not after legal flags them post-launch.

Workflow Overview

Word
Copilot Studio

Step 1: Formalize the rules the agent checks againstCopilot Studio

In Copilot Studio's Knowledge settings, upload the actual brand guidelines and regulatory documents (disclaimer requirements, claims-substantiation rules, restricted-terms lists) as source files.

A verbal summary or a one-off prompt describing the rules will drift and get forgotten between sessions, while a connected knowledge source stays authoritative as those documents get updated.

Step 2: Run every piece of content through the agent before it shipsCopilot Studio

Build this check into the campaign approval flow as a required gate — attach it to whatever step content already passes through before publishing — rather than a link someone has to remember to open; an optional check is the first thing skipped under deadline pressure.

Prompt idea:

Check this campaign copy against our brand guidelines and regulatory requirements. Flag any tone, terminology, or claims-substantiation issue, citing the specific guideline it violates. Give a clear pass/fail.

Step 3: Route failures back with the specific citationWord

Send the content back with the agent's cited guideline attached word-for-word ("violates Section 4.2: all performance claims require a dated source"), not your own paraphrase of the issue — the original citation is what the content owner needs to actually fix the right thing on the first try instead of guessing.

Step 4: Track recurring violation typesWord

Keep a running log of which guideline gets cited most often across campaigns — three different writers all flagged for the same restricted-terms violation points to a training gap worth fixing at the source, not three separate one-off corrections that'll keep recurring.

Check the work

  • Confirm every "pass" actually reflects a real check, not content the agent didn't fully evaluate.
  • Verify flagged claims-substantiation issues against your actual supporting data before removing a claim.
  • Spot-check a "passed" piece manually now and then to confirm the agent isn't missing real issues.

Source: Microsoft Copilot Scenario Library — Marketing (2026)

Expected Outcome

A pass/fail compliance check on campaign content, with specific violations cited against brand or regulatory rules.

✓

AI is the right call here

Formalizing the rules (step 1) and tracking recurring violations (step 4) are process work. Checking tone and claims-substantiation against prose guidelines is real reading comprehension — a restricted-terms list alone would catch only the easiest violations.

Non-AI alternative: A simple find/replace or regex list catches literal restricted terms, but tone and claims-substantiation nuance need the agent's actual reading comprehension.

Related Workflows

WORK WITH LIMINALS

Ready to roll this out beyond one person?

Workflows like this tend to raise real governance and licensing questions once more than one person is using them — that's exactly what we help with.