The count needs context.
A business can appear in an AI answer in several ways. It might be recommended as a provider, linked as the source of an explanation, or used as a caution about service availability. Counting all three as a recommendation would tell the wrong story.
Our Rooter-Man Austin study made this distinction practical. The same business appeared in service shortlists, useful guidance and availability cautions. We reviewed the visible answers alongside the source evidence to understand those different roles.
Three questions behind the number.
| Observation | What it tells you | What to inspect next |
|---|---|---|
| Business mention | The name appeared in the answer. | Was the identity correct? What role did the business play? |
| Answer citation or link | The visible answer referenced a source. | Does that page support the nearby claim? |
| Returned source | The search system exposed a source in its metadata. | Was it actually linked or used in the answer? |
A directory is a lead for investigation.
A recurring directory can be worth reviewing. First confirm whether the business already has a profile, whether its details are accurate and whether that source is relevant to the customer’s question. Seeing a directory in the evidence does not prove that buying a listing will improve AI visibility.
The same principle applies to a missing citation. The business might have a complete listing that did not appear in the sample. A useful recommendation needs both the research observation and a separately verified business gap.
Turn the distinction into better work.
- Show what was observed, with a date and denominator.
- Review identity and context before calling a mention a recommendation.
- Confirm the business facts before making profile or content changes.
- Measure inquiries separately from research visibility.
This gives a business a clearer way to evaluate the work: what changed, why it was selected and what happened after it shipped.
Related study: Where AI finds Rooter-Man Austin.