Citations, mentions and recommendations: three different metrics of AI visibility
Published on September 29, 2026
Metrics and Monitoring
Three clinics can have the same number of appearances in ChatGPT and completely different results. Citation, mention and recommendation are three distinct metrics: citation measures whether the AI uses you as a source, mention measures whether it says your name, and recommendation measures whether it points patients to you. Only the last one fills your schedule.
At Bincoo, while monitoring thousands of clinics, we've seen that most marketing reports confuse the three, and that's why many people think they're doing well in AI when they're only being cited as a technical reference.
More than 10,000 companies already use Bincoo to be recommended by ChatGPT, Gemini and other platforms.
The three metrics explained with real examples
Imagine a patient asks Gemini about teeth whitening in your city and the answer is:
- Citation: "According to the Smile Clinic blog, laser whitening lasts about 2 years." Here you're the source of a fact, not always with a link, but with source recognition.
- Mention: "Among the options in the area are Smile Clinic, OdontoVida and DentalPrime." Your name appears, but with no clear sign of preference.
- Recommendation: "For whitening in [your city], I recommend Smile Clinic for their experience and reviews." Here the AI chose you as the answer to the question.
The gap between the levels is huge in commercial value. A 2026 search behavior study found that AI users click or book with the first recommendation in more than 60% of cases, and almost never look beyond the third option mentioned.
The AI visibility pyramid
| Level | Metric | Question it answers | Commercial value |
|---|---|---|---|
| 3 | Recommendation | Does the AI point patients to me? | Patient at the door |
| 2 | Mention | Does the AI know I exist? | Brand awareness |
| 1 | Citation | Does the AI use me as a source? | Topical authority |
The right strategy climbs the pyramid: first make sure AIs know you exist (mentions), then build authority (citations), and finally win recommendations, which are what pays the bill.
The most common mistakes when measuring visibility
- Adding mentions and recommendations into one number: this hides recommendation drops behind neutral mention spikes.
- Celebrating technical citations: being the source of a statistic doesn't bring patients if nobody recommends your clinic.
- Measuring once and never again: AI answers change weekly, and without recurring measurement you won't see the drop coming.
- Ignoring the tone of the mention: a mention in an answer about complaints counts negative, not positive.
The most reliable signal at each level
In practice, each level of the pyramid has a dominant trigger. Citations respond to content with data and statistics. Mentions respond to consistent presence on portals, directories and reviews. Recommendations respond to a set: specific content about your treatments, visible positive reviews, and mentions in independent sources that AIs trust.
That's why order matters. A clinic that tries to skip steps and only produces commercial content without authority usually sees few recommendations. The base supports the top.
How to measure the three separately
The method is simple, but it demands consistency. Run your market prompts every week and classify each answer: cited as a source? mentioned? recommended? Note the tone and the position. After four weeks, you have a real baseline to decide where to invest content.
A practical example: if your clinic is cited as a source in 20% of prompts, mentioned in 15% and recommended in only 3%, the diagnosis is clear. You already have topical authority; what's missing is turning that authority into referrals. The path is direct commercial content: treatment pages with pricing, reviews, and objective answers to decision-stage questions.
That's exactly what Bincoo does automatically, with a weekly report separating the three metrics and showing the prompts where you're cited but never recommended. All in the Essential plan, at R$ 39.00/month.
Conclusion
- Citation, mention and recommendation measure different stages of AI visibility.
- Only recommendation converts into patients; the other two are steps to get there.
- Measuring the three separately reveals where your strategy is leaking.
If you want to see today which level of the pyramid your clinic is at, access Bincoo and get the weekly diagnosis.
Frequently asked questions
"What's the difference between a citation and a mention in AI?"
A citation is when the AI uses your clinic as the source of a piece of information, saying where it got a fact. A mention is when it simply says your clinic's name in some context. A clinic can be cited often as a source and rarely recommended as an option.
"Is an AI recommendation worth more than a mention?"
Yes, and by a lot. A mention can be neutral or even negative, like appearing in a comparison that favors a competitor. A recommendation is the AI pointing patients to your clinic, which converts directly into patients.
"How do I increase my clinic's recommendations in AI?"
By publishing content that directly answers patient questions, keeping positive reviews on sources AI reads, and ensuring consistent information about your clinic across the web. Bincoo runs this plan weekly starting from the [Essential plan, at R$ 39.00/month](/pt/planos).
"Why did my mention count go up but no patients came?"
You're probably earning neutral or contextual mentions, not recommendations. The metric that connects AI visibility to real patients is the recommendation rate, and that's the one that should drive your decisions.
"Can I be recommended without being cited as a source?"
Yes, and it's common. AIs recommend based on the pattern of mentions and authority they find on the web, without always citing a specific source in the answer. That's why the three metrics need to be measured separately.
