Step 1: Add your brand context
Enter your website, category, product description, key differentiators, and markets you want to study. Avoid confidential personal data. Verify any product or performance claims yourself before including them in a brief.
Step 2: Confirm your own Page
The proposed Core plan includes one primary own-brand Page separately from 20 competitor Pages. Confirm the correct Page identity rather than relying on a name match. A website alone does not establish all associated advertiser Pages.
Step 3: Select competitors
Review research suggestions or add a Page manually. Confirm names and identifiers. Start with direct alternatives and add other references only when they answer a useful creative question.
Step 4: Select supported markets
Choose up to five supported markets per Page in the proposed plan. Only markets actually offered in the live product should be selectable. Each Page-and-market combination has its own coverage status.
Step 5: Review the starting evidence
Existing records can be shown before a new check completes. Look at the observation window and timestamps. A pending competitor does not mean there are no ads; an incomplete historical period must remain labeled as incomplete.
Step 6: Ask your first research question
Try: “What messages and offer structures appear in the available static ads for these competitors? Cite examples and state the gaps.” Review the examples before generating a deeper report.
Step 7: Choose the next test
Convert a supported observation into an original brief. Define what changes between versions and how your own measurement will evaluate it. A competitor’s creative history is context, not a substitute for your results.
Start with one competitor and one question.
You do not need to fill every tracking slot before the first review. Choose one Page that is relevant to your product and one question that would help your next creative decision.
For example: “How does this advertiser explain product quality?” is a more useful starting point than “Analyze everything.”
Confirm the advertiser.
Check the Page identity rather than relying only on a familiar name. Different Pages can represent different markets, storefronts, or businesses. Select a supported market context and inspect the available coverage.
A pending check means the research is not ready for that scope. It does not mean the advertiser has no ads.
Open a small set of relevant records.
Start with a few examples that actually address your question. Read the available copy, inspect the image when accessible, and note the observation dates.
Keep unavailable media visible as a limitation. Do not ask the AI to describe an image it cannot access.
Write one observation before asking for a recommendation.
Describe what you can directly see: a material detail, a product demonstration, a stated offer, or an objection answered in the headline. Keep your interpretation in a separate sentence.
This makes it easier to evaluate whether an AI-generated explanation is supported.
Add your own product context.
Supply the relevant product facts, differentiators, approved claims, and production constraints. Explain what you are trying to learn.
Without that context, even a careful competitor review can produce suggestions that do not fit your business.
Leave with one original test brief.
Choose a clear customer question and a message your product can support. Specify the visual direction, required assets, and the variable you intend to compare.
Save or export only through the options actually available in your workspace. You can also take the brief into your existing planning process.
The first useful outcome is not a large collection. It is a better-defined next step.