Being mentioned by AI and being recommended by it are two different results, and they usually come in that order. Brand mentions on real publications make engines aware of you and teach them what you do; that shows up as a higher mention rate. Recommendation is the next step: being named early, with a reason, when a buyer asks “which one should I choose?” A brand can be well known and still rarely recommended.
This guide explains the difference, how to measure each, and how to tell when you have moved from one problem to the other.
What is the difference between a mention and a recommendation?
Compare two answers to the same illustrative prompt, “What is a good accounting tool for a small design studio?”
In the first, the engine lists eight tools and your brand is seventh, with no comment. In the second, it suggests three tools and puts yours first, adding that it suits small creative teams that bill by project. Both count as a brand mention. Only the second is a recommendation.
| Mentioned | Recommended | |
|---|---|---|
| What it means | The engine knows you and names you | The engine suggests you as a good fit |
| Where you appear | Anywhere in the answer, often in long lists | Near the top, often among few options |
| What comes with it | Your name, maybe a category | A reason tied to the buyer’s need |
| Typical prompts | “What brands make…”, “List tools for…” | “Which should I choose…”, “Best for…” |
| Main metric | Mention rate | Average position and top-three share |
| Main lever | More accurate mentions in context | How you are described, compared and positioned |
Why are mentions the first step?
Because an engine cannot recommend a brand it does not know. Across 75,000 brands, Ahrefs found web mentions were the factor most closely correlated with AI Overview visibility, and that about a quarter of brands had no AI Overview mentions at all. That is correlation, not proof of cause, but it matches how retrieval works: engines build answers from what other sites say, and most sources AI references are not brands’ own sites, according to McKinsey.
Mentions in the right topic context also shape how you are described. If publications consistently place your name next to “small creative teams” or “project billing”, engines have material to connect you to those needs. Our guide on how context shapes AI answers covers this in detail.
Why is being recommended a separate problem?
Because the questions that lead to recommendations are different, and so is the competition. BrightEdge found that commercial wording such as “deals” or “where to buy” produced 4 to 8 times more brand mentions than informational queries, and that 44% of prompts named no brand at all. When a buyer asks a choosing question, the engine names brands readily, but it names a short list, and the order matters.
Even category leaders do not hold the top spot reliably. In SparkToro’s headphone test, the best-known brands appeared in only 55 to 77% of responses. Recommendation is a contest for a few slots in answers that change from run to run, not a fixed ranking you can claim once.
The signals involved also differ. Seer Interactive found that ranking on page one of Google was the strongest predictor of being named by ChatGPT in finance and SaaS questions. That study measured mentions, not recommendations, but it suggests classic search visibility still counts. Comparisons, reviews, use-case pages and how you stack up against named alternatives plausibly feed the reasons an engine gives. Being known is necessary; being chosen depends on how convincingly the web explains why you fit.
How do you measure where you stand?
Use one prompt set built from real buyer questions, and split it in two:
- Discovery prompts. Broad category questions, such as “What companies offer [your category] in [your market]?” These test whether you are known.
- Choosing prompts. Decision questions, such as “Which [category] is best for [specific need]?” or “[Your brand] vs [competitor], which is better for [use case]?” These test whether you are recommended.
Run each prompt several times on each engine, since answers vary. For discovery prompts, record your mention rate. For choosing prompts, record your position in each run, the share of runs where you are in the top three, and whether the engine gives a reason for suggesting you. Track your share of voice against the two or three competitors who keep appearing. Our guide on measuring AI visibility has a full method, and the free AI visibility self-check gives you a scoring grid to start.
How do you know you are ready for step two?
Read your results as one of three patterns:
- Rarely named, even on discovery prompts. You have a visibility problem. Focus on getting known: accurate mentions in context, entity data, crawler access. Recommendation work is premature.
- Named on discovery prompts, low or absent on choosing prompts. Engines know you but do not yet see you as the answer. This is the classic signal you are ready to move from mentions to recommendations.
- Named and often recommended, but behind a competitor. You are already in step two. The work is sharpening how you are compared and which use cases you own.
One check applies to all three: are the facts right? If engines describe you inaccurately, fix that first, because recommending a wrong version of your brand helps nobody. Our guide on fixing AI answers that get your brand wrong covers that.
What does step two involve?
At Odys Blue Ocean we treat this as two services. The Blue Ocean Brand Mentions is step one: getting known, so AI finds, reads about and cites you. Step two is The Blue Ocean GPT, which focuses on getting a brand consistently named near the top of ChatGPT answers for its chosen topics. Brands already well known to AI can start at step two.
Whoever does the work, the honest framing is the same. Nobody controls what an engine says on a given run, and nobody can promise a position by a date. What can be done is to improve the material engines draw on, then measure how often and how high you appear, before and after. Ahrefs found that brands visible on one AI platform tend to be visible on the others too, but each engine still needs checking on its own.
If you are not sure which pattern you fit, you can get a free AI visibility check and see your mention rate and position across the main engines before deciding.
What to do next
- Split your buyer questions into discovery and choosing prompts, and run each set several times on ChatGPT, Gemini and Perplexity.
- Note which competitors take the top spots on choosing prompts, and what reasons the engines give for them.
- If a competitor dominates, read our playbook for when a competitor owns the AI answers in your category.
Frequently asked questions
What is the difference between being mentioned and being recommended by AI?
A mention means the engine names your brand somewhere in an answer, perhaps in a list or as background. A recommendation means it suggests you as a good choice for the buyer's need, usually near the top and with a reason attached. You can be mentioned often and still rarely recommended, which is why the two are measured separately.
How do I know if my brand is ready to focus on AI recommendations?
A useful sign is that engines already name you for your core topics in a fair share of repeated runs, and describe you accurately. If you are rarely named at all, or the facts are wrong, start with getting known. If you are named but sit low in lists, or mostly appear when buyers ask about you by name, you are likely ready for step two.
Can I measure my position in ChatGPT answers?
Yes, as an average across many runs rather than a fixed rank. SparkToro found the same brand list almost never repeats between two runs, so record where you appear in each run of each prompt and average it. Track the share of runs where you are named first or in the top three, and whether the engine gives a reason for suggesting you.
Do brands already well known to AI need brand mentions first?
Not necessarily. If engines already name you often and describe you correctly, more general mentions add less. Such brands can go straight to work on recommendations: how they are compared, which use cases they are tied to, and how they appear in answers to buying questions. A baseline measurement tells you which case you are in.
See what AI says about your brand right now.
Tell us your website, your topic and your market in about a minute. We run a free AI visibility check and show you where you stand.
Get my free AI visibility check →