Being found in answers, not just search results

For twenty years, search visibility meant one thing: rank on Google. The entire SEO industry — the tools, the agencies, the $139/month subscriptions — was built around ten blue links.

The ten blue links are becoming a paragraph. When someone asks ChatGPT, Gemini, Grok, or Perplexity for a recommendation, there is no page two. There's an answer, and your brand is either cited in it or it isn't.

The new metric: Share of Voice

Share of Voice in AI search is simple: across a set of representative customer prompts, how often is your brand mentioned in the answers? Not ranked — mentioned. Cited. Recommended.

This is a different measurement job than SEO. Rankings are deterministic-ish: query, get results, record position. AI answers are generative: the same prompt can produce different answers, different citations, different framings. Measuring visibility means running prompts repeatedly, across engines, and counting citations — a sampling problem, not a lookup.

The engines differ, too. ChatGPT, Gemini, Grok, and Perplexity each have their own retrieval, their own citation behavior, their own blind spots. Winning in one and invisible in another is the common pattern — which is why single-engine tracking lies by omission.

Why the incumbents are awkward here

SEMrush — the deepest SEO platform, starting at $139/month after its May 2026 plan restructure — bolted AI visibility onto legacy plans. It's there, but it's an add-on to a rankings product, and the white-label reporting agencies actually need sits at the $549/month tier.

That's the tell: the incumbents built for the old game and are retrofitting the new one. A tool built for AI answers from the start measures Share of Voice as the core metric, not as a feature on the pricing page.

What agencies actually need

Agencies don't need another dashboard. They need client packs: the visibility story for each client, packaged and presentable — Share of Voice across the four engines, trending over time, which prompts are won and lost, how the client compares to competitors. Under the agency's brand, not the tool's. White-label isn't vanity; it's the agency's business model.

And the methodology has to be inspectable. A visibility number you can't audit — what prompts, how often, how citations are counted — is a number you can't defend to a client. The client pack should document the method, not just the score.

The honest trade-offs

• Measurement isn't optimization. Knowing you're invisible in Gemini's answers doesn't make you visible. The toolkit measures; the content and authority work that moves the number is still yours (or your agency's).
• Sampling has noise. Generative answers vary. Share of Voice is a statistical measure — trends over time are signal, any single day's number is noise. Anyone selling you certainty about AI answers is selling something else.
• The engines change. Citation behavior shifts as the models update. A measurement tool has to re-baseline continuously, and historical comparisons need version awareness — the same discipline as versioned scoring anywhere else.

Being #1 on Google still matters. But the question your customers are asking isn't "what ranks" anymore — it's "what does the answer say." Measure the answers.