72% of B2B Software Buyers Use ChatGPT to Evaluate Vendors — And Half of Brands Are Invisible

In July 2026, MarketScale published a report with a number that should reset every B2B software marketing plan for H2: 72% of B2B software buyers now use ChatGPT to evaluate vendors, and roughly 51% of the brands they consider have zero citations in the LLM's answers.

Translate that into a shortlist meeting: three of four buyers open ChatGPT before they open a Gartner Magic Quadrant. And one out of every two brands they ask about is not in the answer at all — not "ranked lower," not "mentioned briefly." Not there.

Why 51% of brands are invisible

ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews do not rank web pages. They retrieve from a training corpus and a live grounding layer that heavily weights third-party authority — press coverage, analyst reports, structured data, and entity graphs like Wikidata and Crunchbase. Brand websites are almost never the primary source.

Most B2B software PR programs still measure clips and impressions. Those metrics tell you a story was published. They do not tell you whether the story lives in a source LLMs retrieve from, whether the brand is resolved as a distinct entity, or whether the coverage is recent enough to survive the freshness weighting that ChatGPT Search and AI Overviews now apply.

What "showing up" actually requires

Being cited by ChatGPT in a vendor evaluation prompt is a compound signal. It requires:

• Entity resolution. The brand must exist as a distinct, disambiguated entity across Wikidata, Crunchbase, LinkedIn, and schema.org — otherwise the LLM cannot map the query to your company at all.
• Tier-1 and vertical press density. Coverage in the outlets LLMs already trust — Forbes, TechCrunch, WSJ, Axios, Fortune, and the vertical trade publications for your category — is the raw material of citations.
• Structured data and speakable schema. On-page markup that makes quotable claims easy for retrieval systems to extract.
• Freshness. Pages older than ~18 months with no meaningful update get cited significantly less. Quarterly refreshes are now table stakes.
• Measurement. You cannot manage what you do not measure. Track LLM citation share on the prompts your buyers actually type — not just Google rankings.

What Crackle PR is doing about it

Crackle PR runs a dedicated AI Visibility Practice that treats earned media as citation infrastructure for generative AI engines. Every engagement covers entity resolution, tier-1 media relations, GEO-native structured data, and monthly LLM citation tracking through our AI Visibility Index. For a deeper walkthrough of the underlying mechanics, see our guide to GEO & LLM optimization in PR.

The takeaway for B2B tech CMOs

The MarketScale number is not a curiosity — it is a re-baseline. If most of your buyers are shortlisting through ChatGPT and half of the vendors they consider are invisible, the winning move is not more clips. It is a citation program: earned coverage designed to feed the retrieval layer, entity work that makes the brand legible to LLMs, and measurement that tells you whether it is working.

If you want to see where your brand sits on that curve today, talk to Crackle PR.