Why it matters.
Ahrefs ran an experiment on its own AI mentions and reported on 6 July 2026 that only 6% of new mentions came from answers citing its own pages; “the other 94% came from third-party content”. About a quarter of citations appeared once and never returned, and the longest unbroken run was 52 days.
Your own site is necessary but not sufficient. The pages that already discuss your competitors are where new mentions of you are most likely to come from.
The 10-minute self-check.
Pick two competitors the engines name for your category, then search Google for:
"Competitor A" "Competitor B" -"Your company"- Note the first ten results: lists, directories, comparisons, reviews, articles, forums.
- Mark which ones also appear in the engines’ source panels (play 05).
Steps.
- Choose three to five competitors the engines name for your panel questions.
- Find pages naming at least two of them: search operators, a backlink tool’s link-intersect report, and the engines’ source panels.
- Classify each page and its fix: directory (claim the profile), list (pitch with evidence), review site (ask clients), article (offer data or comment), forum (take part).
- Rank the pages: cited by an engine first, then top-ten rankings, then the rest.
- Work five a month, and log each with its date and outcome.
- Re-run the audit every quarter.
This audit is the work queue behind our AI search visibility programmes.
What we saw on prolinkage.com.
How to measure it.
- Gap size: pages naming two or more competitors and not you, quarterly.
- Closed gaps: pages that now name you, with dates.
- Outcome: share of panel runs that name you (play 04).
Sources.
- Ahrefs, self-promotional content AI SEO experiment, 6 July 2026.
- Prolinkage AI self-test, sources panels (Perplexity, Google AI Overview), 6 November 2026.
- Prolinkage change ledger, directory profile corrections (Clutch, DesignRush), 7 November 2026.