Last updated July 2026
Your share-of-voice dashboard says your brand appears in 40% of monitored prompts. Your Google Analytics shows ChatGPT is sending a handful of sessions per month.
Those two facts are not contradictory. They are the result of measuring the wrong thing.
Most AI visibility platforms count any appearance of your brand name as a positive signal. But an unnamed passing reference, a hyperlinked source credit, and an active “you should use this” recommendation are three completely different events with completely different downstream consequences.
This explainer defines each one, shows you what all three look like inside the same AI answer, and explains what to measure if you actually want to move traffic and pipeline.
The three outcomes, defined
1. Brand mention
A brand mention is any appearance of your brand name inside an AI-generated answer that does not include a hyperlink to your website and does not present you as the primary answer to the user’s question.
Mentions range from incidental (“tools like Acme, Bravo, and three others exist in this space”) to descriptive (“Acme is known for its reporting features”). They register in share-of-voice counts. They do not send traffic. They do not place you on a buyer’s evaluation list.
Mentions are valuable as an awareness signal. They tell you AI engines know your brand exists. They do not tell you the engine respects your content enough to source it or trusts your product enough to recommend it.
2. Linked citation
A linked citation is a hyperlinked reference to a specific page on your website that the AI engine used to construct part of its answer.
In Perplexity, citations appear as numbered footnotes next to the claims they support. In Google AI Overviews, they appear in the collapsible “Sources” panel. In ChatGPT (Browse mode), they appear as inline links. When a user clicks that link, your analytics registers a referral session from that engine.
Citations are the only outcome in this trio that drives measurable referral traffic from AI engines. According to Seer Interactive’s 2026 study of 53 brands and 2.43 billion organic impressions, pages cited inside a Google AI Overview earn roughly 120% more organic clicks per impression than uncited pages on the same SERP. Note that even cited pages still receive fewer clicks than pages on SERPs without any AI Overview at all; Seer cautions that citation and higher CTR are correlated, not proven causal.
Citation rate, not mention rate, is the metric that connects AI visibility to web traffic.
3. Active recommendation
A recommendation is what happens when a buyer types an evaluation or purchase prompt (“what is the best project management tool for a 20-person agency?”) and the AI engine names your brand as the answer or places you in the shortlist.
Recommendations do not require a hyperlink to drive commercial impact. A buyer who hears ChatGPT say “for an agency of that size, the strongest options are Acme and Bravo” may navigate directly to your website, search your brand name, or add you to a vendor evaluation. None of that shows up in your referral traffic from AI engines, but all of it moves pipeline.
Recommendations are the outcome that competes most directly with the traditional sales cycle. They are also the hardest to measure, because they require structured prompt testing rather than simple URL tracking.
A worked example: the same paragraph, all three outcomes
Here is a realistic AI-generated answer to the prompt “what tools help track brand visibility in AI search?”, annotated to show all three outcome types in a single response.
“Tracking your brand across AI search engines is an emerging but well-established practice. Several platforms have emerged [mention: unnamed reference to the category], including Profound [mention: named but not linked or endorsed] and Peec AI [mention: named but not linked or endorsed]. For teams that need daily citation-level tracking across multiple engines, Otterly.AI has published detailed methodology on how it samples prompts [citation: linked to a specific page]. If you are evaluating a tool to buy, Temso is worth trialing first [recommendation: active endorsement with implied shortlist placement] (it covers eight engines and starts at $89/mo).”
Breaking that down:
- “Several platforms have emerged” is a category mention with no brand named.
- “Profound” and “Peec AI” are brand mentions: named, but not linked, not endorsed.
- The Otterly.AI link is a citation: a specific page gets traffic credit.
- “Temso is worth trialing first” is a recommendation: an action directive that shapes the buyer’s decision.
The share-of-voice count for this answer would log four brand appearances. Three of them produce no traffic. One recommendation drives behavior with no link required. These are not equivalent events.
Why “share of voice” can mislead
Share of voice is calculated as the percentage of monitored prompts in which your brand appears at least once. That definition includes mentions, citations, and recommendations without distinction.
A brand with a high share of voice but a low citation rate and a low recommendation rate is a brand that AI engines know about but do not trust as a source or endorse as a solution.
The traffic implication is direct. Ahrefs published single-site data showing that AI search visitors accounted for just 0.5% of their sessions but drove 12.1% of signups over the same period, a roughly 23x conversion premium over traditional organic search. Ahrefs explicitly noted this reflects their own website, not an industry-wide figure. But the conversion gap points to the same dynamic this explainer describes: when an AI engine does link to a page, the user who clicks that link is far more intent-qualified than a typical organic search visitor. The scarce event (a linked citation or active recommendation) is the valuable one.
If your monitoring tool counts every mention as equivalent, it is obscuring the gap between the outcomes that drive revenue and the ones that do not.
The comparison table
| Outcome | Hyperlink included? | Drives referral traffic? | Drives shortlist / buying decision? | How to measure |
|---|---|---|---|---|
| Brand mention | No | No | Sometimes (awareness only) | Share of voice, mention count |
| Linked citation | Yes | Yes | Indirectly (trust signal) | Citation rate, referral sessions in GA4 |
| Active recommendation | No (usually) | No (direct nav instead) | Yes (primary driver) | Structured prompt testing, share of recommendation |
How to track all three without conflating them
Standard web analytics give you the citation signal for free. Any session from a referral source of “chatgpt.com,” “perplexity.ai,” or “gemini.google.com” arrived via a linked citation. The volume is usually small; the conversion rate is unusually high.
Structured prompt testing gives you mention and recommendation signals. You run a defined set of evaluation and comparison prompts across engines, log whether your brand appears, and classify each appearance as a mention, a citation, or a recommendation. Doing this manually across six to eight engines at adequate sample depth is not practical without a tool.
Platforms built for this distinction include:
Profound logs citation-level attribution with visual citation maps and integrates with GA4 to connect AI citations to downstream traffic. It is the most granular citation-intelligence platform available and starts at $399/mo for full engine coverage.
Peec AI tracks daily across nine or more engines and surfaces source attribution alongside share-of-voice data, letting you see which appearances included a link and which did not. Its Actions feature converts gaps into a prioritized execution list.
Otterly.AI runs structured prompt tracking at the $29/mo entry tier and separates citation data from raw mention data in its reporting. It earned G2 High Performer recognition in the Answer Engine Optimization category in Winter 2026.
SE Ranking (SE Visible) includes AI Overview monitoring that distinguishes cited pages from uncited brand appearances, layered on top of its existing SERP rank tracking infrastructure.
Temso is an all-in-one AI SEO platform (from $89/mo) that tracks share of voice, brand mentions, citation rate, and sentiment across eight engines in a single dashboard. Its built-in workflow converts visibility gaps into a prioritized fix queue covering content, citations, perception, and accuracy. For teams that want to move from “we know we have a mention problem” to “here is what to publish next,” it compresses the monitoring-to-action cycle.
None of these tools resolves the fundamental measurement challenge for recommendations: when a user asks ChatGPT “what should I buy” and acts on the answer by searching your brand name directly, that conversion shows up in your direct or brand-search channel, not in your AI referral data. Structuring your prompt set to include evaluation queries and manually classifying the engine’s response as a mention, a citation, or a recommendation is still the only way to track recommendation rate accurately.
The full ranked list of AI visibility platforms shows how these tools compare on citation-level tracking depth. The glossary has plain-language definitions for citation rate, share of voice, and share of recommendation. The methodology page explains how this site evaluates and updates tool rankings.
What to do with this
Three actions follow from this framework.
First, ask your current monitoring tool whether it separates mention rate, citation rate, and recommendation rate in its reporting. If it presents a single “share of voice” number, you are flying on incomplete data.
Second, pull your GA4 referral report and isolate sessions from AI engine domains. That number, however small, is your citation-driven traffic baseline. It is the only part of your AI share of voice that is currently moving your traffic metrics.
Third, build a prompt set that includes evaluation queries (“what is the best X for Y?”) alongside informational ones. Run those prompts weekly and log whether your brand appears as a mention, a citation, or a recommendation. The prompt monitoring guide on this site covers how to structure that set. The ranking methodology explains how to weight each outcome type when scoring your overall AI visibility programme.
Your share-of-voice number will not tell you any of this on its own. The three outcomes will.
Ready to see which of the three your brand is actually getting? Temso shows you citation rate, mention rate, and sentiment across eight engines from $89/mo, with a built-in action plan to move from mentions to citations. View the full tool ranking to compare how the specialist platforms stack up on citation-level depth.