Set up prompts and markets

Before you start

Step by step

1. Collect candidates from product attributes. Pro and above can also use connected reviews and contexts. Search volume is a demand proxy, not actual AI prompt volume.
2. Separate unbranded discovery questions from questions containing your brand or product name. Choose prompts you can act on instead of tracking every candidate.
3. Specify the question, country, language, and supported engine in the tracking screen. Check remaining prompt and market limits before saving.
4. Review the meaning of each question and its first collected answers. Consolidate overly broad or duplicate prompts.
5. All plans track every three days by default. Enterprise contracts can adjust the interval.

Check the result

You should have a defined prompt set and market conditions by product family. Interpret different language, country, and engine combinations separately.

If something goes wrong

• If answers are missing, check collection status and last refresh time.
• Distinguish active tracked prompts from saved research candidates when managing limits.
• Do not interpret a percentage change from a small sample as a major outcome.

Related guides

• Product and customer contexts
• Read AI answers and results
• Scale workflows across products and markets