What are AI brand mentions? Definition, types and how to track them

AI brand mentions are brand names in the text of AI answers. Learn what counts as a mention, the 5 mention types, and how to measure and track them over time.

  • Alvaro Peña de Luna Alvaro Peña de Luna
  • date icon

    Wednesday, Sep 30, 2026

AI brand mentions are occurrences of a brand’s name (a company, product, service, app or website) in the text of an answer generated by an AI engine such as ChatGPT, Perplexity or Google AI Overviews.

They are the raw material of every AI visibility metric. How often you are mentioned, how early, in what tone and next to which competitors tells you how AI engines present your brand to people who ask about your category.

What counts as an AI brand mention

A precise definition avoids inflated or missed counts:

  • It is in the text. The brand is named in the answer itself, in plain text or as the text of a link that presents the brand.
  • It is not a URL or a source. A domain inside a URL, or a page listed in the sources, is a citation, not a mention. A source attribution (“according to…”) and an indirect reference (“the company”) do not count either.
  • It goes by identity, not exact spelling. A short or slightly different form of the name still refers to the same brand. A translated name is a different string and is usually treated as a different name.
  • Every occurrence counts. A brand named three times in one answer has three mentions.

Types of AI brand mentions

Counting mentions is not enough, because the way a brand is mentioned changes its value. Mencoro classifies each mention by five types:

  • Recommendation: the answer endorses the brand.
  • Comparison: the brand is compared with named alternatives.
  • Listing: the brand is an item in a list or table.
  • Example: the brand illustrates a concept.
  • Reference: a factual mention.

Each mention also has a tone (positive, neutral or negative), a qualifier (direct, or conditional when it is limited to an audience or use case, such as “for large teams”) and a position (its place among all the brands the answer names, where 1 is the first). Type describes structure and tone describes judgment: advice to avoid a brand is a recommendation with a negative tone.

Worked example. An answer says: “For small agencies, Acme is the best choice. Beta is cheaper than Acme but harder to set up. Other options are Gamma and Delta.” Acme gets a direct, positive recommendation in position 1 and a second mention inside the comparison. Beta gets a comparison in position 2 with a mixed judgment. Gamma and Delta get neutral listings in positions 3 and 4. Four brands, five mentions, and very different value for each.

Why AI brand mentions matter

An AI answer usually names a handful of options. The brands named are the ones the user considers, and the way they are named (recommended, compared, listed) frames the decision before any website visit. Mentions can also come from sources you do not control, such as review sites and forums, so they reflect your reputation as much as your own content.

How to measure AI brand mentions

Mentions feed several metrics. Two of the most common are:

Mention rate = answers that mention the brand ÷ answers analysed × 100

Favorability = (positive mentions + ½ × neutral mentions) ÷ all mentions × 100

For example, 12 positive, 6 neutral and 2 negative mentions give a favorability of (12 + 3) ÷ 20 × 100 = 75. Read it with your mention rate: a high favorability on very few mentions is a thin result. To compare against competitors, weight the mentions by type and tone and calculate your share of voice.

Common mistakes

  • Counting only exact spellings of the brand name and missing its variants.
  • Treating a citation of your domain as a mention.
  • Counting mentions without looking at their type and tone.
  • Ignoring the untracked brands that engines name next to you.
  • Reading one answer as a verdict: answers vary from run to run.

How Mencoro detects AI brand mentions

In every answer from ChatGPT, Perplexity, Google AI Overview and Google AI Mode, a language model finds each brand named and classifies it by owner, type, qualifier, tone and position, and each mention is checked against the visible text. The owner is your brand, a tracked competitor or an untracked competitor (another brand offering the same thing). Untracked competitors appear in each answer but count in no metric until you track them. Mentions then feed Coverage, Favorability, Share of voice and Mention position. Detection accuracy is measured against real answers reviewed and corrected by a person. See How Mencoro works for the full rules and the AI brand monitoring feature for how it looks in the app.

FAQ

Frequently asked questions

No. A mention is your brand named in the text of the answer. A link in the sources is a citation, even when it points to your domain. The two are measured separately because they tell you different things.
No. Being recommended is worth more than being one item in a long list or a passing factual reference, and a positive mention is worth more than a negative one. That is why mentions are classified by type, tone and qualifier instead of only counted.
Engines build answers from the sources they retrieve and from what the model learned in training. If the pages that rank and get cited for your category name your competitors and not you, the answer usually follows. Look at which sources are cited for those prompts to find the gap.
You can run your prompts by hand in each engine and log the results in a spreadsheet, or use a free one-off checker. For trends over time, several engines and several competitors, a tracking tool saves you the manual work and keeps the method consistent.

Start tracking your brand in AI search today

Monitor how AI engines cite your brand, track keyword positions, and benchmark against competitors, all in one platform.