Insights

AI answers quote people, not pages.

About 40% of B2B buyers now start vendor research with an AI tool. What the assistants cite decides who exists in the answer.

Infographic comparing what AI assistants cite: 75% of cited LinkedIn content comes from individual profiles against 25% from company pages; half the people quoted have under 10,000 followers.
In short

Meltwater analysed 9.5 million AI citations across six assistants. 75% of the LinkedIn content cited came from individual profiles and 25% from company pages. Half the people quoted have under 10,000 followers, and 72% of citations are original writing rather than reshares. With about 40% of B2B buyers starting vendor research in AI tools and only around 4% clicking through to sources, the job for leadership comms is to be quotable under a real name.

Where buyers now start

Ask ChatGPT about your sector, then read the answer and look for your leadership. More buyers start there than most comms teams assume: about 40% of B2B buyers now begin vendor research with an AI tool, roughly level with traditional search for the first time. The useful question is no longer where your leaders rank in a results page. It is whether the model quotes them at all.

What gets cited

Meltwater looked at 9.5 million AI citations across six assistants over four weeks. Of the LinkedIn content that got cited, 75% came from individual profiles and 25% from company pages. Follower count barely moved the result: half the cited creators had under 10,000 followers, and 72% of what got quoted was original writing, not reshares.

The quote is the impression

Almost nobody clicks through. The Reuters Institute found only about 4% of people often follow an AI answer back to its source, against 19% from search. So the citation is the whole encounter: the buyer reads the answer, absorbs who is in it, and moves on. The company page never gets the visit.

How we think about it

The models lift specific, original expertise, published under a real name, in plain language, often enough to be present when someone asks. That is a production discipline, not a growth hack, and it favours substance over volume. Not more corporate posts, clearer ones, from a person. We treat a leader’s public record as the raw material the assistants will quote, and we build it deliberately.

We wrote about the ownership side of this problem in Nobody owns AI visibility yet.

Want to know how your leaders appear in AI answers?

We run a fixed-scope AI Visibility Audit: your leaders, your buyers' questions, across the assistants they actually use.

Sources
  • Meltwater, LinkedIn AI Visibility Study: 9.5 million AI citations across six assistants, 2026.
  • 6sense, B2B buyer research, 2026.
  • Reuters Institute, Digital News Report 2026.