People Are the New Homepage: What 9.5 Million AI Citations Say About B2B Visibility
Ask ChatGPT or Google's AI Mode to recommend a service provider — an agency, a SaaS tool, a consultant — and read the answer carefully. The sources it leans on are often not the polished websites those companies spent months building. Increasingly, they are posts and profiles written by individual people.
That is the headline finding of a study from media intelligence firm Meltwater, which analysed 9.5 million citations generated by AI platforms across B2B categories.
Why AI visibility suddenly matters this much
AI answers are not a novelty layer on top of search anymore; they are replacing a meaningful share of clicks. Pew Research Center's analysis of real browsing behaviour found that when Google shows an AI summary, users click a traditional result on only 8% of visits, versus 15% when no summary appears — and click a link inside the summary itself on just 1%. That's the same structural shift we covered in the zero-click piece, seen from the citation side.
In other words: for a growing slice of your prospects, the AI's answer is the shortlist. If your firm isn't in it, you were never considered.
What the study found
1. LinkedIn punches far above its weight
Across those 9.5 million citations, LinkedIn was the second-most cited domain overall at 0.53%, behind only YouTube at 1.52%. Note the absolute numbers are small; most citations still go to the long tail of the open web. But among platforms a B2B marketer can directly publish on, LinkedIn sits near the top.
2. AI models cite people, not company pages
This is the finding that should change your content plan: 75% of LinkedIn citations pointed to individual member profiles and their content. Only 25% came from company pages. More than half came from members with under 10,000 followers, which suggests models are rewarding clarity and usefulness rather than audience size.
That inverts how most businesses allocate effort. The typical setup — a carefully managed company page posting announcements while founders and senior staff post occasionally, if at all — is optimised for exactly the kind of content models pass over.
3. Decision-stage, structured content wins
The most-cited content shares a recognisable anatomy: clear formatting with bullets and numbered lists, strong headings, named companies and tools, and real numbers. The formats that dominate are unglamorous and specific — "best X" listicles, side-by-side comparisons, "how to choose" guides.
What this means for your business
The practical takeaway is not "post more on LinkedIn." It is more precise than that.
Put real names on your expertise
If your founder, design director or head of delivery has opinions worth paying for, some of them should be published under their own name. A company page lends legitimacy — keep it active — but the citations flow to people. For a services business, two or three genuinely knowledgeable voices posting consistently will likely out-earn the brand account for AI visibility.
Write the content buyers ask AI for
Prospects ask comparison questions: best CRM for a 20-person sales team, Webflow vs custom development, how to choose a UX agency. The cited content answers exactly those questions — with structure, named options and honest trade-offs. Yes, that sometimes means naming alternatives to your own service. Comparison content that only ever concludes "us" reads as advertising to humans and, apparently, to models.
Structure for machines without writing like one
The anatomy of cited content — headings, bullets, specific names, real numbers — is simply well-organised writing. You don't need to stuff keywords or write robotically. You need to be extractable: a model should be able to lift a clean, attributable claim from your post. It's the same discipline that wins citations anywhere else.
Refresh on a rhythm
Recency counts, so a small library of evergreen comparison pieces updated quarterly will beat a large archive of stale ones.
The caveats worth keeping
Two honest qualifications. First, citation rates are low in absolute terms — LinkedIn's 0.53% share means your website, your reviews and third-party coverage still carry most of the weight. AI visibility is a layer on top of a sound SEO and content foundation, not a replacement for it.
Second, this is a single study, run by a vendor with a product in this space, using its own methodology. Treat the percentages as directional rather than settled. The direction, though, is consistent with Pew's click data and the broader shift to AI-assisted research.
Where to start this quarter
A realistic 90-day play: pick two people in your business with real expertise; have each publish one structured, decision-stage piece per fortnight — a comparison, a "how to choose", a costed breakdown; anchor every piece with specifics, tools named and numbers included; and keep the company page publishing the proof (case studies, work, announcements) that gives those personal posts institutional weight.
None of this is complicated. It is mostly the discipline of writing useful things in public, consistently — which has quietly become a ranking factor for the answer engines your next client is already using.
Founder of Plutinos Technologies, a Noida design and development studio working with startups and growing businesses across India since 2016. Writes about UX, accessibility, web performance and how search is changing. Led the 2026 audit of 105 Indian SMB websites.
More from this author →Where does your brand show up in AI answers?
Content, SEO and design from one team — starting with an honest read on your current visibility.