Being Mentioned in AI Answers Is Not the Same as Being Recommended
The article argues that current AI visibility tools incorrectly equate brand mentions with recommendations, failing to distinguish between being advised, merely listed, or dismissed. To address this, the author implemented a low-cost classifier that categorizes AI responses into four distinct states: recommended, listed only, dismissed, and absent. Testing on a demo corpus revealed that high mention counts often mask a lack of actual endorsements, with many answers naming brands without steering users toward a specific choice.
- ▪A classifier was applied to 63 stored AI answers to categorize brand mentions into recommended, listed, dismissed, or absent states.
- ▪The analysis cost approximately $0.0033 for 252 judgments, demonstrating that detailed semantic analysis is economically viable.
- ▪In the demo corpus, the most-mentioned brand was named in 100% of answers but actively recommended in only 3.2% of them.
- ▪The system uses a decision model that returns probabilities rather than prose to ensure consistent and countable results without parsing errors.
- ▪Mention counts are criticized for treating dismissals and neutral listings as equivalent to positive recommendations, which can mislead marketing strategies.
Hacker News (AI / LLM) files mainly under ai. We currently carry 6,606 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Story provenance
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | CueScout |
| Canonical URL | https://cuescout.com/blog/mentioned-is-not-recommended |
| Publication time | Sun, 27 Sep 2026 10:35:43 +0000 |
| Retrieval time | 2026-09-27T10:50:41.120Z |
| Last seen | 2026-09-27T10:50:41.120Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | FG1SHtxaf0qa · 1 stories |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
The short versionEvery AI visibility tool counts mentions, ours included. A mention count gives the same score to an answer that recommends you and an answer that names you in a list and then advises someone else.There are four things being named can mean: recommended, listed only, dismissed, absent. A mention count can only see the fourth one.We put a classifier on 63 stored answers from ChatGPT, Perplexity and Gemini: 252 judgments in 7.3 seconds for $0.0033, about $13 per million judgments. The cost is the point, because reading everything a second way stops being a project.In that run the most-mentioned brand was named in 100% of answers and advised in 3.2% of them, and 52.4% of answers recommended nobody at all.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at CueScout.
Discussion
0 commentsMore from CueScout
