If you want to know whether ChatGPT, Perplexity, Gemini, Copilot, or Google AI experiences mention your brand, you do not need to begin with an expensive monitoring platform. You can start with a fixed prompt set and a spreadsheet.
This free starter benchmark gives you 20 buyer-oriented prompts to reuse across AI platforms. Copy the prompts into a sheet, replace the bracketed placeholders, and record the evidence from each run.
How to use the benchmark
- Replace [category], [audience], [problem], [brand], and [competitor] with terms that fit your business.
- Use the same prompt wording each time you rerun the benchmark.
- Run the prompts on the AI platforms that matter to your audience.
- Record mentions, recommendations, citations, competitors, source URLs, and notes.
- Repeat on a consistent schedule so you can distinguish patterns from one-off variation.
20 AI-search buyer prompts to track
| # | Prompt group | Prompt template | What to watch |
|---|---|---|---|
| 1 | Category discovery | What are good [category] tools for [audience]? | Which brands appear at all? |
| 2 | Category discovery | Which [category] solutions are best suited to [use case]? | Recommendation position and reasoning |
| 3 | Category discovery | What are affordable [category] options for [audience]? | Price/value positioning |
| 4 | Category discovery | Which [category] tools are easiest to use for a small team? | Ease-of-use claims and cited evidence |
| 5 | Category discovery | What [category] products should a beginner consider? | Beginner-oriented recommendations |
| 6 | Problem solving | How can I solve [problem] without [undesired tradeoff]? | Whether your category is proposed as a solution |
| 7 | Problem solving | What tool can help me [desired outcome]? | Brand and product mentions |
| 8 | Problem solving | What is a practical way for [audience] to handle [problem]? | Recommended workflow and sources |
| 9 | Problem solving | How should I evaluate [problem] before making changes? | Methods, frameworks, and cited domains |
| 10 | Problem solving | What mistakes should I avoid when trying to [desired outcome]? | Recurring pain points and source citations |
| 11 | Comparison | What are good alternatives to [competitor]? | Whether your brand appears as an alternative |
| 12 | Comparison | How does [brand] compare with [competitor] for [use case]? | Strengths, weaknesses, and factual accuracy |
| 13 | Comparison | Which is better for [audience]: [brand] or [competitor]? | Decision criteria used by the model |
| 14 | Comparison | What are the main differences between leading [category] options? | Which brands define the comparison set |
| 15 | Comparison | Which [category] option offers the strongest value for [use case]? | Value claims, price references, and citations |
| 16 | Purchase intent | Which [category] tool should I choose if I care most about [priority]? | Recommendation and justification |
| 17 | Purchase intent | What [category] tools are worth paying for for [audience]? | Paid-product recommendations |
| 18 | Purchase intent | What should I look for before buying a [category] product? | Selection criteria you may need to address on your site |
| 19 | Brand audit | What is [brand] known for? | Brand narrative, factual accuracy, and sources |
| 20 | Brand audit | Who competes with [brand], and how are they different? | Competitor set and category positioning |
What to record for every answer
At minimum, keep these fields:
- Date
- Prompt ID
- Prompt text
- AI platform
- Brand mentioned? Yes/No
- Brand recommended? Yes/No
- Brand cited? Yes/No
- Competitors mentioned
- Source URLs or cited domains
- Short evidence note
Do not collapse everything into one score
ChatGPT, Gemini, Perplexity, Copilot, and Google AI experiences can return different brand and source patterns for the same topic. Track platform-level results first. A single combined score can be useful as a summary, but it should not hide the underlying evidence.
How often should you run it?
Weekly is a practical default for a small benchmark. Daily checks are more useful during a launch or short-lived test; monthly checks can be enough for slow-moving categories. Consistency matters more than constantly changing the prompt set.
When the starter benchmark is no longer enough
Move beyond the starter set when you need more prompt coverage, repeated runs, competitor share-of-voice, source tracking, action prioritization, or a reusable dashboard. At that point, a structured workbook can save a lot of manual cleanup.
The AI Search Visibility Evidence Tracker expands this starter method into a reusable Excel system with fixed prompt sets, per-run evidence, mentions, recommendations, citations, competitor tracking, source tracking, repeatability, charts, and an action queue.
For the full methodology behind the benchmark, read How to Track AI Search Visibility Without Relying on a Mystery Score.
Methodology references
- Semrush: How to decide which AI search prompts to track
- Semrush AI Visibility Index methodology
- Ahrefs: Custom prompt tracking guidance
This benchmark is a structured sampling method. AI-generated answers vary and the results do not represent every private user conversation or guarantee mentions, citations, traffic, leads, or sales.