LearnGrokUse this pack to turn scattered campaign inputs into a decision that can be defended before creative work starts. It is for marketing managers and brand strategists who need to choose a message territory, not merely generate options.
The pack works in sequence. Start with evidence, compare the possible positions, recommend one territory, then test the claim boundaries and challenge the decision.
Key point
Build the decision from evidence
A distinct message is not enough. It must match a real audience tension, product proof and a competitor gap.
1. Gather the four inputs
Collect the latest versions of these documents before using the first prompt:
- Audience research, including interview findings, survey summaries, search behaviour or sales objections.
- Current messaging, including the campaign brief, website copy, sales deck and brand platform.
- Competitor claims, ideally copied from ads, landing pages, decks or a dated competitor scan.
- Product evidence, including feature documentation, validated outcomes, case material and known limitations.
Keep the source labels when pasting material. A statement without a source may still be useful context, but it cannot carry a campaign promise. If the documents are long, paste the relevant passages rather than an unlabelled summary.
Use Evidence intake and gap register first. Its job is to separate what is known from what the team assumes. It also identifies conflicts, such as research saying buyers want simplicity while current messaging leads with technical depth.
Watch out
Do not treat competitor copy as proof
A competitor can claim a benefit without delivering it. Record the claim as a market signal, not as a verified product fact.
2. Set up message territories worth comparing
Before running Positioning matrix, write three or four candidate territories. A territory is a strategic promise, not a headline. For example, it should identify a customer tension and a product-backed outcome, rather than rely on words such as “better”, “simple” or “trusted”.
Include the existing direction as one candidate. This makes the approval discussion honest. The team can see whether the proposed new direction is genuinely stronger or merely newer.
Run the matrix with the evidence intake attached. Review the cells marked unproven before discussing scores. A territory with high distinctiveness but low credibility is usually a research or product decision, not a copywriting problem.
Check
The matrix is usable when
Every recommended score has a source reference, and every missing proof point has an owner. If either is absent, the matrix is opinion dressed as analysis.
3. Make one recommendation
Use Recommended message territory after the matrix. Give it the campaign objective and likely channels, such as paid social, search, email or a landing page. Channel context matters because a broad brand promise may not survive a short ad format, while a narrow performance message may not support a wider launch.
Take the resulting positioning statement to the approval meeting, not a page of potential taglines. The statement should make four things plain:
- Who the campaign is trying to move.
- The tension or situation it addresses.
- The benefit the product can credibly promise.
- The proof that makes the promise believable.
The approval conditions are as important as the recommendation. They show what must be true before the team uses the territory externally.
4. Check claims before creative work
Run Claim and proof review whenever the territory includes outcomes, comparisons, speed, cost or category-leadership language. It does not replace review by the qualified internal person. It gives that person a traceable list of phrases, proof and unresolved questions.
Avoid approving a direction because the central phrase “sounds right”. Ask whether the supplied evidence supports the meaning an ordinary prospect is likely to take from it. Keep the approved evidence-backed elements with the creative brief.
If you need to check current platform behaviour or documentation before preparing the workflow, use the xAI documentation overview. Details can be version-dependent.
5. Challenge the result in the room
Use Approval challenge memo as a final pre-read or run it during the working session. Share the objections alongside the recommendation. Do not hide the rejected alternative. It tells stakeholders that the team considered the nearest credible route and made a deliberate trade-off.
| If you see this | Do this next | Do not do this |
|---|---|---|
| A strong promise with weak proof | Assign an evidence owner and hold the phrase | Use softer copy to conceal the gap |
| Clear proof but heavy competitor overlap | Find a narrower audience, moment or benefit | Claim the competitor is weak without evidence |
| Conflicting research findings | State the conflict and commission a focused check | Select the finding that suits the preferred direction |
| No territory is credible enough | Pause approval and reduce the promise | Ask creative to solve a product evidence problem |
Stop
Do not brief final ads from an unapproved claim
Creative variation cannot make an unsupported promise defensible.
How to tell the output is wrong
Look for certainty that exceeds the inputs. Warning signs include a competitor weakness with no source, a customer motivation that was never researched, a product outcome stated as fact when the material only describes a feature, or a high score with no evidence reference.
Also check for false distinction. A territory can use different words while making the same promise as every competitor. Read the campaign promise column without the brand name. If it could sit on a competitor landing page unchanged, the territory is not yet distinct enough.
When the pack does not work
If the first output is mostly not supplied or unproven, stop refining prompts. Get the missing product proof, audience evidence or competitor examples. If the matrix produces several tied options, do not force a winner. Use the stated decision gaps to define a short research or stakeholder task, then rerun the relevant prompt with the new evidence.
When the recommendation is challenged successfully, return to the matrix rather than rewriting the conclusion. The work is only complete when the approved territory, its proof and its boundaries agree.
