Wikipedia, LinkedIn, Press: What AI Chatbots Cite About Schools

Why LLMs almost never read your website directly

LLMs don't crawl your site in real time before answering a prospect's question — they draw on a fixed, pre-selected corpus dominated by a handful of trusted third-party sources. Your admissions page, your programme brochure, your carefully worded "About Us" copy: none of it carries much weight unless it has already been echoed somewhere the model trusts more than you.

This is the part of Generative Engine Optimisation (GEO) that trips up most marketing teams. GEO, as Search Engine Land defines it, is the practice of positioning a brand to be cited, recommended, or mentioned inside AI-generated answers — and citation depends on what other sources say about you, not just what you say about yourself.

A recent 2026 study by 5W Research, distributed via PR Newswire, found that Wikipedia and Reddit together drive over 25% of ChatGPT citations in the US. Traditional broadsheet authority barely registers by comparison: the Wall Street Journal, the New York Times and Bloomberg do not appear in the top 20 cited sources at all. For a school, that means the institutional gatekeepers you might expect to matter — legacy press mastheads — are not where the citation weight sits.

Our own GEO pillar guide for schools covers the technical side: Schema.org markup, entity density, freshness. This article covers the other half — the off-site reputation signals that determine whether an AI engine trusts your institution enough to name it.

Wikipedia: the page you can't buy

A Wikipedia entry is one of the strongest single citation triggers available to a school, but it cannot be created or edited the way a brochure page can — it has to be earned through independent coverage and built by someone other than the institution itself. Profound's 2026 citation-pattern analysis found that Wikipedia alone accounts for 7.8% of all ChatGPT citations, and nearly half — 47.9% — of citations within ChatGPT's top 10 cited sources overall.

Wikipedia's notability guideline for organisations is explicit on what qualifies an institution for an entry. It requires significant coverage in reliable secondary sources independent of the subject — a press release, sponsored content, or a quote from your own comms team never counts toward that bar. The article has to be built from journalism, academic citations, or regulatory coverage that exists whether or not the institution asked for it.

This is why editing your own institution's page directly is a risk, not a shortcut. Wikipedia's volunteer editors actively screen for conflict-of-interest editing — undisclosed edits from an institutional IP range, or a marketing team polishing its own entry, are routinely flagged, reverted and sometimes escalated to a permanent COI notice on the article's talk page. That notice is publicly visible and does lasting reputational damage.

The correct sequence runs the other way:

• Earn independent coverage first. A regional newspaper profile, a sector publication feature, a QAA report mention, or an academic citation gives editors something citable to build from.
• Propose the page through the right channel. Use Articles for Creation, or raise the case on the article's talk page, rather than creating or editing the entry directly.
• Let a neutral editor assess notability. If the coverage is there, an uninvolved editor will normally accept the draft; if it isn't, no amount of internal pressure will substitute for it.
• Keep the page current through primary sources you don't control. Once live, update requests should point to press coverage or official records, never to internal talking points.

The practical implication is sequencing: press coverage has to come before a Wikipedia entry can exist, which is why the next two sections matter as much as the encyclopaedia page itself.

LinkedIn: the identity LLMs treat as verified

LinkedIn functions less like a social network and more like an identity register that LLMs treat as verified information about an institution and its people. ALM Corp's 2026 analysis of 325,000 prompts found that LinkedIn is the second most-cited domain across ChatGPT, Gemini, Google AI Overviews, Copilot and Perplexity combined — a striking result for a platform most schools treat as a recruitment channel rather than a GEO asset.

The format matters as much as the presence. The same ALM Corp analysis found that LinkedIn Articles account for 50–66% of cited LinkedIn content depending on the platform, far ahead of ordinary company posts or static company pages. A long-form article written by a named academic or a director carries more citation weight than a promotional update from the institution's own handle.

Profound's data adds a platform-specific nuance: Perplexity, unlike ChatGPT, skews toward LinkedIn, NIH and G2 — community and professional platforms rather than encyclopaedic ones. Perplexity also cites more sources per answer on average (21.9) than ChatGPT (10.4), so a thin LinkedIn presence costs you visibility on Perplexity specifically, not just generically.

Three actions follow directly from this:

• Complete and keep consistent your institution's LinkedIn page. Programme names, accreditations, location and founding year should match your website and Wikipedia entry exactly — inconsistency across sources reduces the model's confidence in any single one.
• Publish LinkedIn Articles under named authors. A head of school or programme director writing 800 words on graduate outcomes or curriculum changes is more citable than the same content posted as a company update.
• Activate personal profiles, not just the institutional page. Individual professional profiles capture roughly 65% of organic reach on LinkedIn, against approximately 5% for company pages — so a director's byline article, shared from their own profile, reaches (and gets indexed by) far more of the platform than the same words posted from the institution's account.

Press and media: the coverage that counts (and the coverage that doesn't)

Independent editorial coverage gets cited by AI engines; a press release sitting on your own newsroom page almost never does, because the model treats self-published announcements as promotional rather than evidentiary. The distinction is not subtle from the model's perspective — one is a claim made by the subject, the other is a claim verified by someone with no stake in the outcome.

That doesn't make press releases useless. A well-targeted release is frequently the mechanism that triggers the article that does get cited. A journalist picking up your release on a new apprenticeship partnership, a TEF Gold renewal, or a graduate outcomes report converts a promotional statement into independent verification — and it's that resulting article, not the release, that becomes citable.

For UK schools, the outlets worth prioritising are the ones that already carry weight in the sector's information ecosystem: The Guardian's education section, Times Higher Education, and the Complete University Guide. Coverage in any of these does double duty — it feeds the AI corpus directly, and it supplies exactly the kind of independent secondary source a Wikipedia entry needs.

A practical filter before issuing anything: ask whether a journalist with no relationship to your institution would consider the story newsworthy on its own terms. Rankings movement, regulatory milestones, verifiable outcome data and named research all clear that bar. Generic "excellence" language does not.

Building your source stack: a 3-step method

Building a citation-worthy source stack is sequential, not parallel — press coverage has to exist before Wikipedia notability can be argued, and both need to be in place before LinkedIn content has independent facts to reference. Treat it as a 90-day build rather than a single campaign.

Step 1 — Secure independent press coverage (weeks 1–4). Identify one verifiable, newsworthy fact — a ranking change, an accreditation renewal, a graduate outcomes figure, a new partnership — and pitch it to a specialist education journalist rather than issuing a generic release. One piece of coverage in a recognised outlet is worth more to your citation profile than ten internal blog posts.

Step 2 — Build the LinkedIn layer (weeks 3–8, overlapping). Once you have an external fact to reference, turn it into a LinkedIn Article authored by the relevant director or academic, and have them share it from their personal profile. Cross-check that your institutional LinkedIn page matches your website and any existing press mentions exactly — name, accreditations, location.

Step 3 — Propose the Wikipedia entry or update (weeks 6–12). With independent press coverage and a consistent LinkedIn presence in place, submit a draft via Articles for Creation, or raise an update request on the article's talk page if a stub already exists. Reference only third-party sources in the draft; never cite your own website as evidence of notability.

This sequence complements two things covered elsewhere. The Schema.org markup that makes your programme pages machine-readable is covered in a separate article in Skolbot's GEO series — it's what makes your site legible once an LLM decides to look. And the full 90-day execution plan, including the technical and measurement steps, is laid out in our 90-day action plan to get cited by ChatGPT and Perplexity. If your priority is the reputation side rather than press — Google reviews, Reddit threads, alumni sentiment — that's covered separately in our 90-day reputation plan for higher education.

Track the impact of this work the same way you'd track any GEO investment: run a monthly set of test prompts across ChatGPT and Perplexity and log whether your institution is named. Our guide on the KPIs and monthly ritual for tracking ChatGPT and Perplexity visibility sets out the exact query set and cadence to use.

Wikipedia vs LinkedIn vs press: a comparison

Two things stand out from this table. First, no single source is sufficient — ChatGPT and Perplexity favour different platforms, so a stack limited to one channel caps your citation ceiling on the other engine. Second, technical markup and off-site sourcing are not substitutes for each other. Schools with structured Schema.org markup achieve an average of +12 percentage points in AI visibility — in the UK specifically, ChatGPT cites schools in 29% of relevant answers and Perplexity in 38%, against a 19% European average (Source: Skolbot GEO Monitoring, 500 queries × 6 countries × 3 AI engines, Feb 2026). Markup gets you found; Wikipedia, LinkedIn and press get you trusted enough to be named with confidence. For the full list of signals that sit alongside these two, see our breakdown of 15 signals LLMs evaluate before recommending a school.

FAQ

Do I need a Wikipedia page to be cited by ChatGPT?

No, but it materially improves your odds. Wikipedia accounts for 7.8% of all ChatGPT citations and 47.9% of citations within its top 10 sources, according to Profound's 2026 analysis — so a well-sourced entry meaningfully raises your citation probability, though schools without one can still be cited via LinkedIn or press coverage.

Can I write or edit my own institution's Wikipedia page?

You shouldn't, and doing so carries real risk. Wikipedia's conflict-of-interest checks routinely catch undisclosed institutional editing, revert the changes, and can leave a permanent COI notice on the talk page. The safer route is proposing content through Articles for Creation or a talk-page request, backed by independent secondary sources.

Why does LinkedIn matter more than our own press releases?

Because LLMs treat LinkedIn as a verified professional record and press releases as self-published promotion. ALM Corp's 2026 analysis found LinkedIn is the second most-cited domain across five major AI platforms, with LinkedIn Articles making up 50–66% of cited content — while a press release with no resulting editorial pickup is rarely cited at all.

Should we prioritise press coverage or LinkedIn first?

Press coverage generally comes first, because it supplies the independent evidence that both a Wikipedia entry and a credible LinkedIn Article can reference. Building LinkedIn content in parallel is reasonable, but a LinkedIn Article citing only internal claims carries less weight than one referencing an external write-up.

Is Schema.org markup enough on its own, without Wikipedia or press?

No. Markup is necessary but not sufficient — it is what makes your site machine-readable, but the +12 percentage point visibility gain it produces still depends on external corroboration existing elsewhere. Schools with strong Schema.org markup and no independent coverage plateau below schools that have built both.

Wikipedia, LinkedIn and press coverage are not marketing extras — they are the trust layer that decides whether an AI engine names your institution or leaves it out of the answer entirely. Building that layer takes months, not days, which is exactly why schools that start now will be the ones ChatGPT and Perplexity are citing by the next admissions cycle.