Track a fixed set of buyer questions on a schedule, record who is cited, and read it alongside referral traffic.
For teams who want evidence, not guesses, about how they appear in AI answers. You will set up a monthly measurement you can keep.
Everyone with a SaaS product now wants to know whether AI tools recommend them. The honest answer is usually that nobody checks properly, because the answers change from run to run and no dashboard is a reliable source of truth. You can still measure it, with a method that is simple, repeatable and honest about its limits.
1Write a fixed set of buyer questions
Write twenty to thirty questions that your buyers ask, from problem questions to vendor comparisons. Mix branded wording, which shows how you appear to people who know you, with unbranded wording, which shows whether strangers find you.
| Stage | Example question |
|---|---|
| Problem | How do I stop losing trial users in the first week? |
| Category | What are the best onboarding tools for SaaS? |
| Comparison | Fernhollow vs a rival for small teams |
| Brand | What does Fernhollow cost, and who is it for? |
2Run each question more than once
AI answers vary between runs, and between tools. A single check proves nothing. Run every question at least three times in each tool your buyers use, on a schedule, and record the date, the tool and the full answer.
3Record what happened
For each run, note three things, and the sources that were cited instead of you.
- Named: does the answer mention you at all?
- Cited: is one of your pages a listed source?
- Linked: does it link to you?
- Cited instead: which domains appear, and which competitors
- Sentiment: how are you described when you are mentioned?
Calculate it per topic. A share that rises over a few months, on a frozen question set, is a better signal than any single answer.
4Add your own data
Two sources in your own tools fill in what the prompts cannot see.
- AI referral visits: create a channel group in analytics for referrers from AI tools, and keep the domain list current
- Self-reported source: add a how did you hear about us question at signup, with AI tools as options
5Read the trend, and act on it
At the end of each month, write one paragraph per topic: rising, flat or falling, and why you think so. Look at who is cited when you are not. Those sources show where to be present, and the topics where you never appear show which pages are missing.
Worked example
A quarter of tracking (illustrative)
A made-up onboarding tool runs twenty-four questions three times in two assistants each month. In month one it is cited in a small share of runs, mostly on branded questions. Cited instead on category questions are two review sites and a rival's guide.
The team improves its comparison page, corrects its listing on one review site and publishes a clearer guide on the category question. By month three its citation share on comparison questions has risen, while the broad category question still cites others, which gives them the next target.
Say it in 30 seconds
The 30-second version
Write a fixed set of buyer questions, run each several times in the AI tools your buyers use, and record whether you are named, cited and linked. Add AI referral visits and a how-did-you-hear question at signup, and read the trend monthly.
Worksheet
AI visibility tracking sheet
The full guide and a worksheet to fill in, as a PDF you can print or share with your team. The worksheet covers:
- Write 20 to 30 buyer questions across the funnel, branded and unbranded
- Freeze the list for a quarter and note the start date
- Choose the AI tools your buyers use and list them
- Decide the number of runs per prompt, at least three
- and 7 more
One email with the PDF. No list, no follow-ups.
