You’re mentioned
Your company or product is named in the answer text.
Reported as: questions where you’re namedHow we measure
How we measure AI visibility, and what it can’t tell you.
Most companies selling “AI visibility” don’t say how they measure it. Here’s exactly what we do, so you can check our work or repeat it yourself.
Step 1
The questions your buyers ask.
We write a fixed set of questions from your category, your buyer, and your competitors: 10 for the free snapshot, 25 for paid plans. The set stays the same from one measurement to the next, so changes mean something.
| Question type | Example |
|---|---|
| Best-of | “Best field service software for small teams” |
| Problem | “How do I stop double-booking field technicians?” |
| Comparison | “Northstar vs. other scheduling tools” |
| Brand | “What does Northstar do?” |
| Evaluation | “Is Northstar good for a 20-person team?” |
You see and approve the question set before we run it.
Step 2
Four assistants, the same way every time.
We ask every question in ChatGPT (with web search), Perplexity, Gemini, and Google’s AI Overviews. Each question runs in a fresh session with no account history, and we record the date, the assistant, and the location the question was asked from.
We save a screenshot of every answer. Each result in your report links to it.
Step 3
What we record.
For each answer, we log four things. Every number in your report comes from these.
Your source is linked
One of your pages, or one of your news placements, is shown as a source for the answer.
Reported as: answers citing youWho else is named
Which of your competitors appear, and in which answers.
Reported as: share of answers vs. competitorsEvery cited URL
All sources each assistant cites, grouped by domain and type: news, review sites, forums, trade publications, roundups, company sites.
Reported as: your citation mapStep 4
Measure again, same questions.
After we publish and, on Newsroom, reach out, we run the identical question set again: at 30 days on Launch, at 30 and 90 days on Authority, and every month on Newsroom.
Your report compares the two runs side by side and lists which cited sources now mention you, alongside every live placement and whether Google has indexed it.
Limits
What any AI measurement can’t tell you.
Answers vary. The same question can get a different answer an hour later, or for a different person. A single run is a snapshot, not a score, so we track the same set over time and look at the trend.
Models change. Assistants update their models and sources without notice. A shift can come from the model, not from anything you or we did.
Correlation, not proof. If citations rise after a placement, we show you the link between them. We can’t see inside the models, so we can’t prove one caused the other.
That’s why we never guarantee citations or rankings. We guarantee the work in your plan and an honest measurement of what changed, including when nothing did.