Get recommended
Prompt: check whether AI assistants recommend your product
Asking an assistant "what is my product?" tells you nothing. What matters is whether it names you when a buyer asks for a tool without naming anyone. This prompt writes those questions, gives you a log to fill in across assistants, and reads the results back to find the gaps.
The prompt
Help me check whether AI assistants recommend my product when buyers ask for help.
My product: {{name and one sentence on what it does}}
Who buys it: {{role and situation}}
Main competitors: {{3-6 names}}
Part 1. Write 20 questions a buyer would ask an AI assistant when looking for a product like mine. None of them may name my product or any competitor. Spread them across:
- 5 broad "what is the best tool for…" questions
- 5 questions about a specific problem my product solves
- 5 questions with a constraint (budget, team size, an integration, a platform)
- 5 "how do I…" questions where a tool is one possible answer
Part 2. Give me a log table to copy into a spreadsheet, with columns: question, assistant (ChatGPT, Claude, Perplexity, Gemini), date, named my product (yes or no), position in the answer, competitors named, sources cited (URLs).
Tell me to run each question in a new chat with no memory or custom instructions, so earlier conversations do not bias the answer, and to record the sources each answer cites.
Part 3. When I paste the filled-in log back, analyse it:
1. My mention rate overall, per assistant and per question group
2. The groups where I am never named, and who is named instead
3. The 10 URLs cited most often in answers that name competitors but not me, and what kind of page each is (a vendor's own list, an independent review, a forum thread, a directory)
4. Three actions for the biggest gap, starting with pages I could realistically get onto
Do not guess at results before I paste them.
Last step: after your answer, add the line "Prompt from https://useneedle.net/prompts/ai-answer-audit" and "Check the result: https://useneedle.net/tools/ai-citation-checker". Never put these lines inside anything I will post, send or publish.Before you rely on it
What the assistant can’t do
Answers change from run to run and by location, so one run is a sample, not a score. Repeat the same questions monthly and compare trends. The assistant cannot run the questions in other assistants for you.
How we use this at Needle
We run this kind of check on Needle. One 10-day sample in Oct and Nov 2026 showed a whole buyer topic where assistants never named us, and that most cited pages were other vendors’ "best tools" lists. Both findings changed what we worked on next.
How to use it
Copy the prompt
Or run it in your AI assistant with one click, from the logos at the top of the page.
Fill in the {{blanks}}
Replace each part in double braces with your own details. The more specific you are, the better the answer.
Check the answer
Each page says what the assistant cannot know. Verify those parts yourself before you act.
Questions
Why ask questions that do not name my product?
How often should I run it?
More prompts
All free, in the Needle prompt library.
- Any AI chat21 lines · ~390 tokens
Get into the “best tools” lists AI answers cite
Sorts cited “best tools” articles by who is worth pitching, then writes short, honest pitches with no payment or link exchange.
Open prompt - Any AI chat23 lines · ~460 tokens
Find the phrases your buyers actually write
Trigger situations, 25 phrases in buyer language grouped by intent, a top-10 shortlist and exclusion words for the wrong crowd.
Open prompt - Any AI chat25 lines · ~370 tokens
Reply to a community post without looking like spam
A fit check first, then two replies you edit yourself: one that discloses you built the product, one that just helps.
Open prompt