What AI discoverability means
AI discoverability is the ability of AI systems — ChatGPT, Claude, Perplexity, Google's AI features — to find, read, and cite information about you when someone asks. When a hiring manager types “who is [your name]” into an AI assistant, the answer is assembled from pages that are crawlable, machine-readable, and trusted. If no such page about you exists, the answer is a shrug — or worse, someone else with your name.
This is a different property from having a social profile. It is closer to how search engines saw the web twenty years ago: a technical bar that most personal web presence today simply does not clear.
How AI systems actually read the web
The mechanics matter, because they are unforgiving:
- —AI crawlers do not execute JavaScript. GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and Meta's crawler fetch raw HTML. Content that renders client-side is invisible to them. Server-rendered pages are a hard requirement, not a nice-to-have.
- —Answer engines lean on search indexes. ChatGPT search and Google's AI features retrieve from ordinary web indexes. If a page isn't indexed, it can't be cited. Standard technical SEO — sitemaps, canonical URLs, clean status codes — is the entry ticket.
- —Structure decides absorption. Studies of AI citations show pages that actually shape answers are substantive and heavily structured — clear headings, lists, direct definitions — while thin pages get skipped or merely name-dropped.
- —Structured data makes you an entity. Schema.org markup (Person, Article, FAQ) turns prose into facts a machine can lift with confidence: name, role, employer, skills, links.
Why LinkedIn can't give you this
LinkedIn is where your professional information lives — and it is walled off from exactly the systems people now ask. Full profiles sit behind login. The robots.txt disallows AI crawlers. The page is a JavaScript application. Each of these alone would be disqualifying; LinkedIn has all three.
The result is an odd inversion: the network you maintain most carefully contributes least to what AI says about you. Whatever scattered mentions exist elsewhere — a conference page, an old blog, a company directory — carry the answer instead. Most professionals have never audited that surface. AI reads it daily.
Becoming citable: the five requirements
1. A canonical page at a stable URL
AI systems favor stable sources. One page, one URL, maintained over time, beats a scatter of profiles. This is the page you want every “who is” answer to resolve to.
2. Server-rendered HTML
The page must arrive as complete HTML. If your site builder renders content with client-side JavaScript, AI crawlers see an empty shell.
3. Schema.org structured data
A Person entity with your name, title, employer, skills (knowsAbout), education and sameAs links lets a machine extract facts instead of guessing from prose.
4. Indexed, and quickly
Sitemap submission and index monitoring are unglamorous and decisive. A page Google doesn't know exists cannot appear in AI Overviews, and a page Bing doesn't know exists is invisible to much of ChatGPT search.
5. Substance
A name and job title get you selected at best. Real projects, real outcomes, specifics an AI can quote — that is what turns a mention into a citation. Write the page you would want read aloud as the answer.
Where factcard fits
factcard packages all five requirements into one page: a profile at your own URL, server-rendered, marked up with Schema.org Person data, published to search engines through sitemaps and IndexNow, with machine-readable endpoints for AI agents. You fill in the facts once — or import your résumé and let AI structure it — and pay $29 one time. No subscription, nothing to maintain.
If you'd rather assemble it yourself, the checklist above is the whole recipe — it is work, but it is doable. The point either way: AI discoverability is now part of a professional identity, and defaults don't provide it.
Related reading: How to get found by recruiters in the AI era.
Frequently asked questions
Can ChatGPT see my LinkedIn profile?
Generally no. LinkedIn blocks AI crawlers in its robots.txt and walls full profiles behind login. When someone asks an AI assistant about you, your LinkedIn is usually not part of the answer — the AI works from whatever public, crawlable pages exist about you.
What is generative engine optimization (GEO)?
GEO is the practice of making content easy for AI answer engines (ChatGPT search, Perplexity, Google AI Overviews) to retrieve, understand, and cite — the AI-era counterpart to SEO. In practice it means server-rendered pages, clear structure, structured data, and substantive content that directly answers questions.
How long does it take for AI systems to find a new page?
Search-connected assistants can surface a page as soon as it is indexed by the underlying search engine — typically days after submission via a sitemap. Training-data crawlers like GPTBot visit on their own schedule, often weeks. A stable URL that accumulates history beats anything transient.
Do I need an llms.txt file to be AI-discoverable?
No. Google has stated it does not use llms.txt, and log studies show AI crawlers almost never fetch them. What actually matters is server-rendered HTML, structured data, and being indexed in ordinary search — the pipelines AI answers are actually built on.
Is a personal website enough?
Only if it is machine-readable: rendered on the server (not client-side JavaScript), marked up with Schema.org Person data, indexed, and kept current. Most personal sites fail at least one of these. That gap is exactly what factcard exists to close.