Brand share of voice
across every AI engine.

Share of voice in AI search is the percentage of cited sources, across your tracked prompt set, that point at your domain versus the competitor domains you track. Attrifast computes it from the actual citation lists ChatGPT, Claude, Gemini, and Perplexity return for your monitored prompts: your citation count divided by the combined count for you plus your declared competitor domains (up to eight on Pro, three on Starter), recomputed live every time you open the dashboard. Unlike classic SEO share of voice — which is computed from estimated SERP impressions — AI share of voice is a presence metric, not an impression estimate, and it should always be reported alongside the revenue those engines produced.

Classic share of voice (Ahrefs, Semrush, SimilarWeb) is computed from estimated keyword impressions: total search volume across your keywords, weighted by your rank position. It is an upstream proxy for traffic. Brand share of voice in ChatGPT (and the other AI engines) is downstream: it counts whether your brand or domain actually appears in the answer the user sees. The first metric estimates how often you could be seen on a SERP; the second measures how often you are actually named inside an AI answer. They overlap roughly 60-70% in my own measurement, the same way backlinks and AI citations overlap, and the divergence zone is where teams burn the most effort.

Because a 50% share of voice on prompts that send no buyers is worse than a 10% share on prompts that send paying trials. I have watched a SaaS founder grind a Perplexity share-of-voice number from 8% to 22% on a prompt set that turned out to be almost entirely research-stage queries. The score moved; revenue did not. The fix was not "get more citations" — it was "scan a different prompt set," which only became visible once revenue per citation was reported alongside share. That is the loop Attrifast was built to close.

ChatGPT (with web search), Claude, Google Gemini, and Perplexity — all four are scanned on every run. Each engine carries its own sub-share so you can see whether your share is concentrated in one model or distributed. This matters because each engine has different citation behavior: ChatGPT with web search links out reliably and weights brand familiarity in its training corpus, Gemini leans on Google's index, Perplexity always links but favors answer-shaped, freshly updated pages, and Claude often summarizes without a clickable link at all. A single aggregate number hides those differences; the engine-level breakdown is what you actually act on.

You bring the prompt list and tag intent (buying, comparison, research). We seed it for you during onboarding using your existing GSC keyword data and your category, but the list is yours and editable. The share calculation is then run only on the prompts you mark as commercially relevant — so the number reflects buyer queries, not vanity queries. The transparent prompt list is also why the score is auditable: if your team disputes a movement, you can drill into the exact prompt, the exact engine output, and the parsed citation list.

When a visitor lands from chatgpt.com, perplexity.ai, gemini.google.com, or any other AI engine, Attrifast captures the referrer server-side on the first visit and stores a first-party session id. When that session later completes a Stripe checkout, the payment_intent.succeeded webhook fires and we join the payment to the originating AI engine. Aggregate that join across a month and you get share of revenue per engine, alongside share of voice per engine. The two numbers usually disagree — which is the entire point.

Yes. Every scan parses the full citation list each engine returns, not just whether you appeared. So the dashboard shows your share alongside each named competitor's share — you declare up to eight competitor domains on Pro, three on Starter, and can swap them at any time. We cannot tell you the revenue your competitors earned from those citations — only your own Stripe data joins back — but the citation-share leaderboard is fully reported.

Profound and Loamly both publish brand share of voice in AI answers; that is essentially the category they pioneered. The Attrifast difference is the revenue join. Profound is sold by demo into Fortune 500 marketing teams and prices accordingly; Loamly's deeper Intelligence reports start at hundreds to thousands of dollars per report. Attrifast packages the citation-share scan together with the Stripe-side revenue attribution in one product, from $9.99/mo on Starter — the $49/mo Pro plan adds sentiment, answer position, topics, and gap analysis on top. We are not the deepest AI-search analytics platform — we are the only one that ties citation share to actual paid trials at SMB pricing.