AI calls almost every brand 'excellent.' AI visibility is owning the word it won't share.

AI used positive language for almost every brand in Parse's 30-day sample. Across 719,860 descriptors attached to 67,013 brands, 85.3% were positive and only 4.7% were negative. Positive sentiment alone does not separate a brand. The useful question is whether AI consistently connects the brand to a specific attribute that competitors do not share.

AI praises nearly everyone

Parse extracts the descriptors AI uses about a brand, the adjectives and short positioning phrases that appear in answers, and labels each one positive, neutral, or negative. In the most recent 30-day window that produced 719,860 descriptor instances across 67,013 brands. The polarity split is lopsided: 85.3% positive, 10.0% neutral, 4.7% negative, and a rounding error of mixed. Put differently, only 12.6% of brands picked up a single negative descriptor in a month of monitoring. AI is a generous narrator.

Parse sees the same pattern in recommendations, where AI uses positive framing more often than cited sources support. Positive sentiment is therefore a weak standalone signal. The more useful questions are which attributes AI uses, how often they recur, and whether competitors receive the same descriptions.

The praise is generic

The words AI reaches for most are not descriptions, they are ratings. The single most common descriptor across the brand universe was "recommended," applied to 5,671 different brands, followed by "excellent" (4,223 brands), "best" (4,119), "strong" (3,113), "comprehensive," "good," and "popular." These are verdicts. They tell a buyer that the model approves, and nothing about what the product is or who it is for.

The vocabulary is also surprisingly flat. The ten most common descriptors account for only 7.6% of all descriptor instances, the top 50 for 18.0%, and the top 100 for 23.5%. There is no small set of power-words doing most of the work; instead AI spreads a thin layer of interchangeable approval across tens of thousands of brands. And it does not say much per brand: the median brand collected just 3 distinct descriptors in the window, with an average of 8.8 and a 90th percentile of 15. For most companies, the entire AI characterization is a handful of generic compliments.

That is why "AI says good things about us" is the wrong thing to track. It is true of almost everyone, it is shallow, and it does not differentiate. A monitoring number that moves with the tide tells you nothing about your position in it.

Specific descriptions separate brands

Of the 277,963 distinct descriptors AI used, 85.4% (237,411) applied to exactly one brand. Specific and unusual descriptions were therefore far more likely to be brand-specific than the common positive terms. These descriptions are more useful for competitive positioning.

You can see it in the descriptors well-known brands actually own. In the same window, Linear's most-used descriptor was "fast" (231 mentions), Notion's was "flexible" (239), ClickUp's was "highly customizable" (194), Stripe's was "developer-friendly," Snowflake's was "scalable," Mailchimp's was "user-friendly," and Zapier's was "simple." Each of these is a position, not a rating. It tells a buyer what the product is for. But notice what is missing from that list: plenty of equally prominent brands top out at praise. Asana, HubSpot, Slack, Datadog, and Vercel all had "excellent" as their single most-used descriptor, and Shopify's was "best." Those companies are visible to AI, but AI has not yet attached a distinct idea to them.

The ownership math makes the gap concrete. Take the share of a descriptor's total mentions held by its single most-associated brand. For generic praise, that share is tiny, because the word is everywhere: "best" 1.0%, "popular" 1.0%, "recommended" 1.3%, "good" 1.3%, "excellent" 1.4%. No one owns "best." But for specific attributes the top brand holds a real chunk: "fast" 11.0% (Linear), "customizable" 6.1% (monday.com), "lightweight" 4.8% (Notion), "open-source" 4.8% (SigNoz), "enterprise-grade" 2.6%, "affordable" 2.0%. A specific word is 5 to 11 times more concentrated than a generic one. That concentration is the asset. It is the difference between being one of four thousand brands AI calls "best" and being the brand AI calls "fast."

Ranked by how concentrated each descriptor is on its single top brand, the contrast between verdicts and attributes is sharp:

Every attribute clears every verdict: the most owned generic word ("excellent," 1.4%) is still less concentrated than the least owned attribute on this list, so the line between a word you can own and one you cannot is the line between attribute and verdict.

If you want to see how AI engines describe your own brand, run a free brand check — it takes a minute.

What this means for your AI visibility

The practical read is to stop measuring sentiment and start measuring descriptor ownership. Three moves follow from the data.

First, do not use positive sentiment as the main success metric. Positive terms made up 85.3% of all descriptors, so a positive average provides little competitive information. Track which specific descriptors attach to the brand and how exclusively AI uses them.

Second, identify a specific, non-evaluative descriptor AI already uses for the brand and check whether competitors share it. A good target describes an attribute, already has some association with the brand, and is not concentrated in a rival. This provides more information than a larger count of generic positive mentions, just as being named differs from being recommended.

Third, defend it in the evidence. AI assigns descriptors from what it reads. If you want to own "fast" or "secure" or "developer-friendly," the sources AI cites for your category need to say it about you, repeatedly and specifically, so the model has something concrete to repeat instead of falling back on "excellent." Which sources those are is category-specific, as we found in every industry's different AI citation recipe: the pages that can plant "fast" on a finance brand are not the ones that do it in consumer goods.

How Parse maps this to your brand

Parse tracks AI visibility across ChatGPT and Google AI Mode, covering a public index of more than 4.7 million AI responses, 603,000 brands, and 57 million citations. The descriptor layer in this study is the same one Parse surfaces for your brand specifically: the exact words AI uses about you, how positive they are, and, critically, how many competitors share each one. The highest-yield finding is almost never your sentiment score. It is a specific descriptor a competitor owns and you do not, or a specific word you weakly hold that no one has locked down yet. For the source side of that work, our study of which review site ChatGPT trusts most shows where the evidence AI repeats actually comes from. See how AI describes your brand.

How we measured this

This study draws on Parse's first-party descriptor layer, built from AI answers collected across ChatGPT search and Google AI Mode. We used the latest complete rolling 30-day window, May 27 to June 26, 2026. Parse extracts every adjective and short positioning phrase AI attaches to a brand, labels each as positive, neutral, or negative, and aggregates them. In this window that produced 719,860 descriptor instances across 603,556 brand-descriptor pairs, 277,963 distinct descriptors, and 67,013 brands. The grain is one row per brand, descriptor, polarity, and descriptor type, with a mention count attached. Polarity shares sum mention counts across rows. Vocabulary concentration is the cumulative share of total mentions held by the top-N descriptors. Long-tail share is the fraction of distinct descriptors carried by exactly one brand. Descriptor ownership is the top brand's mentions divided by that descriptor's total mentions. Two caveats: this is a single point-in-time window, not a trend, and because these brands are not niche-mapped, ownership is reported globally rather than per category.