SERPTool · Docs

AI Mentions and AI search volume

A third of search traffic now routes through AI surfaces — Google's AI Overviews, ChatGPT search, Perplexity, and friends. Traditional keyword tools are blind to this. SERPTool's AI Mentions feature measures two things per keyword: how often an AI actually answers the query, and which domains get cited when it does.

What the numbers mean

Total AI mentions An aggregate count of times this keyword produced a cited response across tracked AI platforms (ChatGPT, Google AI Overview, and peers). A value like 8,400 means the AI systems surfaced an answer-with-citations for this query roughly 8,400 times in the measurement window.

AI search volume DataForSEO's estimate of how many monthly searches for this keyword happen via AI interfaces rather than classic Google. Useful for comparing: a keyword with 10,000 classic monthly searches and 2,000 AI searches has 20% of its demand moving outside traditional SERPs.

Cited domains A ranked list of the top 20 domains the AI systems cite in their answers for this keyword. These are the effective ranks in an AI-first world — if ChatGPT keeps citing wikipedia.org and nih.gov for your medical query, those are your real competitors for AI-surface visibility.

Why it matters

Classic SEO strategy says: rank in the top 10 organic, get clicks. AI-first strategy says: get cited by the AI even if you're not in the top 10. The citation is the click — the user rarely visits your site, but your content informed the answer they read, and you got brand exposure plus the opportunity to be cited in future answers on related queries.

Three practical implications:

  1. Citation-worthy structure beats rank-worthy structure. AI systems prefer clear definitions, explicit comparisons, direct answers, numerical specifics, and authoritative sourcing. Pages written in that style get cited disproportionately to their organic rank.

  2. Being cited by Wikipedia or a similar high-authority source has outsized downstream effect. AI models use Wikipedia heavily. A citation there cascades into AI answer citations.

  3. Early-mover advantage is real but narrow. AI training cutoffs mean today's content can still reach the next model generation. Later, established sources entrench harder than they do in classic SEO.

When to fetch AI Mentions

AI Mentions costs 40 credits per lookup (vs 1 for standard keyword analysis) because DataForSEO charges per-query and the underlying data is expensive. It's opt-in: there's a button on every keyword detail page that triggers the fetch.

Good candidates:

  • Informational keywords where AI surfaces are displacing organic clicks. High AI search volume = you'll lose classic-SEO clicks; get cited to stay visible.
  • Category-defining queries where being listed in the "what is X" answer drives long-term brand positioning.
  • Comparison keywords (vs, best, alternatives) — AI systems frequently quote specific products and sources; being in that list is high-value.

Poor candidates:

  • Pure transactional keywords (buy, price, near me) — AI surfaces rarely handle these; classic SEO still wins.
  • Brand queries — trivially "you're the answer", no lookup value.
  • Extreme long-tail — AI mention volume is too low to signal meaningfully.

What to do with the data

If your domain appears in the top cited domains: you're winning. Sustain the structural patterns that got you there.

If a direct competitor dominates: analyze their cited pages. What structure, sourcing, or specificity patterns are they using that AI systems reward? Emulate those patterns on your own content.

If the cited list is mostly authoritative neutral sources (Wikipedia, .gov, .edu): the AI treats this as a "settled fact" domain where commercial sources don't rank. Your goal shifts from "be cited for the core query" to "be cited for adjacent commercial queries" (e.g. product comparisons that cite the core definition).

Caveats

  • AI models update. A leader today can drop when the next model ships.
  • Citation patterns leak training preferences, not user click behavior. A heavily-cited page might not drive traffic in the classic sense — but it drives AI-answer quality that shapes future searches.
  • The number is directional, not exact. Treat it as signal, not ground truth.

Next steps