Short answer: ChatGPT, Perplexity, and Google AI Overviews cite the pages that are (1) crawlable by their bots, (2) recognised as a real entity in structured knowledge sources, and (3) written so a direct answer can be lifted in one sentence. Miss any of the three and you stay invisible — no matter how well you rank on Google. The checklist below covers the nine signals that decide it, in priority order, plus a free tool that scans your site for every signal in about 20 seconds.
Google Ranking vs AI Citation: Not the Same Game
Most businesses still think of "search" as a single channel. It is now two channels with different rules. Here is the cleanest way to see the difference:
Google ranking signals reward relevance, backlinks, on-page SEO, page experience, and freshness — they decide whose link appears on the results page.
AI citation signals add bot access (GPTBot, ClaudeBot, PerplexityBot, Google-Extended), entity verification (Knowledge Graph, Wikidata, NAP consistency), structured data (FAQPage, HowTo, Article, Organization), answer-first writing, and citation-friendly formatting — they decide whose sentence gets quoted inside an AI answer.
You can rank #1 on Google and never get mentioned in ChatGPT. The reverse is also true. Modern AI search programs optimise for both — that is the work covered under AEO and GEO.
The 9-Signal AI Citation Checklist
Work through these in order. The first three are non-negotiable — without them the remaining six cannot help you.
1. AI bot access in robots.txt
AI crawlers identify themselves with specific user-agents: GPTBot (OpenAI / ChatGPT), ClaudeBot and anthropic-ai (Anthropic / Claude), PerplexityBot, and Google-Extended (the user-agent Google uses to decide what feeds Gemini and AI Overviews). If any of these are blocked in your robots.txt, those engines cannot quote you. Most accidental blocks come from copy-pasted templates or a CDN's default rules.
Check your file at yourdomain.com/robots.txt. Allow the bots you want to be cited by. If you have content you genuinely do not want included in AI training, scope the disallow narrowly — block /private/, not the whole site.
2. llms.txt at your site root
llms.txt is a short markdown file at yourdomain.com/llms.txt that tells AI models which pages are your canonical, citation-worthy sources. It is not yet a formal standard, but Anthropic, Perplexity, and a growing list of crawlers respect it. For a services business, a useful llms.txt lists your service pages, pricing, locations, FAQs, and a one-line description of each. It takes about 30 minutes once and is one of the highest-leverage early-mover plays in 2026.
3. Structured data: Organization, FAQPage, HowTo, Article
Schema is how you tell AI engines, in machine-readable form, what each page is and what the answers are. The four that matter most for citations:
Organization on the homepage (name, URL, logo, sameAs links to LinkedIn, Wikidata, Google Business Profile) — this is the entity grounding signal. FAQPage on any page with question-and-answer blocks — AI engines extract these directly. HowTo on procedural content. Article (or BlogPosting) with datePublished, dateModified, and an explicit author or publisher — freshness and provenance both matter.
Validate with Google's Rich Results Test or Schema.org's validator. Implemented-but-invalid schema is worse than no schema.
4. Wikidata and Knowledge Graph entity grounding
AI models cross-reference structured knowledge sources before deciding whether a brand is a "real" entity worth quoting. A complete Wikidata entry — founding date, headquarters, founders, official URL, industry classification — is one of the highest-leverage one-time tasks for AI visibility. Pair it with a claimed and complete Google Business Profile and (where eligible) a Google Knowledge Panel claim. This is the core of GEO work.
5. Answer-first writing (the first 100 words)
AI engines preferentially cite pages that answer the question up front. Lead with a 40–80 word direct answer, then expand. The post you are reading does this in the first paragraph — that is the pattern. Burying the answer five paragraphs down hands the citation to a competitor whose page leads with it.
6. Comparison and definition tables
Tables earn disproportionate AI citations versus prose because they are structurally unambiguous — an AI can extract a row and quote it cleanly. Any "X vs Y", "pricing tiers", or "feature comparison" content on your site should be a real HTML table, not a styled div.
7. Author and E-E-A-T signals
AI engines weight provenance. A blog post with a real author byline, a linked author page, and credentials is more citable than an anonymous one. For business content, the equivalent is a clear publisher with verifiable history — date founded, location, leadership. ThinkProfits has operated continuously from Vancouver since 1996 and has served 3,500+ clients across Canada and the U.S.; we list founder, address, and contact on every page because every one of those facts is an AI-resolvable entity signal.
8. Citation-friendly anchor patterns
Use descriptive internal anchor text ("our AEO services", not "click here"), include source attributions for any statistics, and link to authoritative external sources where relevant. AI models follow link graphs to verify claims — a page that links to nothing looks isolated; a page that cites and is cited looks like a node in a real knowledge graph.
9. Freshness signals
AI engines favour recently updated content for time-sensitive queries. Set an honest dateModified in your Article schema whenever you genuinely update a page. Do not fake it — engines compare schema dates against actual content diff, and inflated freshness signals get discounted.
Why You Rank #1 on Google but ChatGPT Never Mentions You
If you are well-ranked but invisible to AI engines, the cause is almost always one of three things. One: a bot in your robots.txt is blocking the crawlers you want quoting you. Two: your entity is not verified — no Wikidata entry, no Knowledge Panel, inconsistent NAP across directories, so the model cannot confidently say who you are. Three: your pages do not lead with answers — the content is good, but it is not structured in a way an AI can extract a sentence from.
Our free SEO audit tool checks every one of those nine signals against your live site in about 20 seconds — bot access, llms.txt, schema, entity signals, answer-first patterns — and shows you exactly where the gaps are. It is the same scan our team runs at the start of every AEO and GEO engagement.
Suggested 30-Day Sequence
Week 1 — Access & entity. Open AI bot access in robots.txt. Publish a basic llms.txt. Claim/complete Google Business Profile. Create or update Wikidata entry.
Week 2 — Schema. Implement (or fix) Organization schema on the homepage, FAQPage schema on every page with Q&A, Article schema on the blog. Validate everything.
Week 3 — Rewrite the top 5 pages answer-first. Pick the five pages you most want quoted (usually homepage, top service pages, pricing). Put a 40–80 word direct answer in the first paragraph. Convert "X vs Y" content to tables.
Week 4 — Monitor. Re-run the audit. Spot-check ChatGPT, Perplexity, and Google AI Overviews on 10 priority queries before and after. Document what changed.
What to Read Next
Use these in order to go from "I understand the checklist" to "I know exactly what my business should do":
- AEO Services → The content-layer playbook — Q&A page architecture, FAQPage / HowTo / Speakable schema, featured snippet and PAA targeting, and weekly answer-share monitoring across ChatGPT, Perplexity, and Google AI Overviews.
- GEO Services → The authority-layer playbook — Knowledge Panel claim and optimisation, Wikidata entity work, AI Grounding Optimisation, and citation acquisition that makes AI engines trust your brand enough to quote you in the first place.
- AEO vs GEO: Which Does Your Business Need? → The plain-English explainer of how AEO and GEO differ and which to prioritise first.
- Free SEO Audit Tool → Scan your site against the nine signals in this post in about 20 seconds.
Stop guessing. Start getting cited.
See exactly where ChatGPT, Perplexity & AI Overviews quote you
Book a free 30-minute consultation. We'll run your business through ChatGPT, Perplexity, Gemini, and Google AI Overviews and show you — live — which of the nine signals above are costing you citations right now.
Book My Free Consultation →30 minutes · No obligation · Vancouver-based, serving North America
