Why AI Cites a Person Instead of Your Company

A scene that keeps repeating: you ask AI a technical question about your industry, and the answer carries a line like "a technical lead at one company said…". You look that company up, and the sentence appears nowhere on its website.

The conclusion first: when AI decides whether a statement is worth trusting, it looks at who said it. A company is an organisation, and an organisation does not speak. A named person can be checked and traced.

But the boundary comes first: this is not an invitation to invent experts. Fabricating author credentials is explicitly targeted under the published rules of several major engines, and the consequences are worse than having no byline at all. That is covered in detail below.

What the data shows: personal profiles beat company pages

Start with the evidence at the level of what is observed.

• One analysis covering 9.5 million AI citations found that of LinkedIn-related citations, roughly 75% came from personal profiles and only about 25% from company pages; around 51% of the citations came from members with fewer than 10,000 followers.
• A second analysis, covering 1.4 billion citations, reports that LinkedIn has become the most-cited domain in AI search; within it, the share of citations going to posts and articles rose from 26.9% to 34.9% in three months, overtaking company pages.
• The same 9.5-million analysis adds a combined figure: LinkedIn, Reddit and YouTube together account for 47.5% of AI citations, while company pages account for only 18.7%.

Scope note: all three figures come from third-party analyses of overseas platform samples, not official data. Platform structure and content ecosystems differ substantially from those in China, so they cannot be mapped directly onto domestic platforms. They indicate a direction — personal expression is credited more readily than an organisation's self-description — not precise ratios.

Why: AI has to decide who is speaking

There is support for this in Google's own documentation.

Google's Search Quality Rater Guidelines explicitly require checking who authored the content, and Google repeatedly sets out a "Who / How / Why" framing — who wrote it, how it was written, and why. Google also states plainly that E-E-A-T is not itself a direct ranking factor, but a framework used when assessing content quality.

One update from September 2025 deserves more attention: AI-generated content is not automatically penalised, but it raises the credibility bar — and pages that pass themselves off as human-written, or that invent author credentials, are rated at the lowest level.

Scope note: the above describes how Google frames content-quality assessment; it is not a promise that a byline earns citations. A byline does not automatically produce citations, but without a traceable author one layer of credibility evidence is simply missing.

A subtler layer: being named is not the same as being cited

An analysis of more than 540,000 answers introduced the term "ghost citations": when a brand is mentioned in the answer, the citation rate for its content is about 53.1%; when the brand is absent from the answer, the rate is about 10.6%. More telling still, roughly 62% of AI citations never produce a brand mention at all.

In plain terms: AI may be citing your content heavily without naming you. It reads your judgement, your data, your experience, and then writes them into the answer as industry common knowledge.

Scope note: this is a third-party statistical analysis and an observational finding; it does not establish causation (being mentioned does not cause a higher citation rate). It does suggest one thing: content alone is not enough — you also need an identifiable entity, and people count as entities.

That is where the human layer earns its keep: a person with a name, a track record and a verifiable identity is harder to absorb into the background than an anonymous page.

What B2B companies should actually do

No need to package anyone as an industry celebrity. Just put the people who genuinely exist in front of the reader:

1. Put a byline on the content, and use real names. Technical articles under the technical lead, export experience under the export manager, product breakdowns under the product engineer. Name, title and a real track record — all three, none optional.
2. Have those people answer real questions in professional settings. Substantive answers in industry communities, expert Q&A and technical forums are read as professional expression far more readily than a company press release. The precondition is real knowledge and a real point of view — not posting for the sake of posting.
3. Your own site needs pages about people. An author page or team page should state what this person is responsible for, what they have done, and what background can be verified. And that information must match what appears on external platforms — name, title and company cannot differ from place to place.

The third point is the one most often skipped: plenty of company sites explain the company on an About page and never mention a single individual. AI goes looking for who is speaking and finds nowhere to land.

The red line: never invent an expert

This deserves its own section, because it is where this line of work goes wrong most often.

Fabricated author credentials are rated at the lowest level under Google's published rules. Inventing a "ten-year industry expert" does not earn credit; it stamps the whole page with the worst trust signal available.

There is also a spillover effect: once one author is judged fabricated, other content on the same site is re-examined too. On bylines, choose fewer and plainer over invented.

Practical advice: two real people beat twenty invented "experts". What a real person writes may be imperfect, but it can be checked — and that is precisely what AI needs.

Three common questions

We are a B2B factory and the boss won't appear publicly. What then?

It does not have to be the boss. A technical lead, a production supervisor, an export manager, a QC engineer — anyone who genuinely knows the business will do. Buyers do not care how big the title is; they care whether the person talks like an insider.

Do bylines create risk — competitors poaching staff, buyers contacting individuals directly?

That is a real concern and each company has to weigh it. The middle path: real name and title, but no personal mobile number or private email; all contact routes point to the company's business channels. You keep a traceable human without fully exposing one.

Do we have to be on LinkedIn?

Not necessarily. Those figures reflect overseas platforms. Build where your buyers ask questions. If your customers are mainly in China, industry communities, expert Q&A and technical forums may fit better than LinkedIn. The point is matching the channel, not following the crowd.

Where this leaves you

To pull this together: AI cites a specific person because it needs a speaker it can verify.

An organisation is abstract; a person is concrete. The people inside your company who actually know the technology, the customers and the industry are already your best sources — except their words have been rewritten as "the company states", which erases the most valuable part.

Start with one small thing: open the last three articles on your site and look for author bylines. If there are none, add real names and titles — the cheapest step available, and the one most often skipped.

Yihang Network (Shenzhen) previously worked in cross-border B2B and Alibaba.com-related services, and has been extending into GEO (Generative Engine Optimization) since 2024. Our view: GEO is not only about explaining the company clearly, but also about explaining the real people inside it.

Data and effect boundaries: the analyses covering 9.5 million, 1.4 billion and 540,000-plus answers are third-party statistical research — observational data drawn mainly from overseas platforms, not officially verified, and not a basis for causal claims. Google's statements on the Search Quality Rater Guidelines, the Who/How/Why framing, E-E-A-T and fabricated author credentials come from its official documentation and update notes. None of the above represents the real market or the results of any specific company. The methods described raise the probability of being recognised and credited; they do not guarantee that a company or an individual will appear in any AI answer, and they promise no timeline or business result.

Sources

1. Third-party analysis: share of personal profiles versus company pages across 9.5 million AI citations (overseas platform sample)
2. Third-party analysis: LinkedIn's share of AI citations and the shift toward posts and articles across 1.4 billion citations
3. Third-party analysis: "ghost citations" and the gap between brand mentions and content citation rates (540,000-plus answers)
4. Google Search Central: Search Quality Rater Guidelines (requirements relating to author checks)
5. Google: guidance on E-E-A-T and content-quality assessment (including the September 2025 update on fabricated author credentials)