Getting cited inside ChatGPT and Perplexity: a working playbook.

Somewhere between a third and a half of the questions your customers used to type into Google now go to ChatGPT, Perplexity, or Claude first. Getting cited in those answers has become its own discipline — adjacent to SEO but not the same thing.

This is the working playbook we use with clients. It changes — the answer engines change constantly — and we update this piece quarterly. What follows is what we believe in February 2026.

Two caveats up front. First: we treat classical SEO as the floor, not the alternative. The brands that get cited in AI answers are almost always the ones already ranking on Google. Skip the fundamentals and the rest of this list won't save you. Second: chasing AI citations as a primary KPI is premature for most brands. Use it as a leading indicator. The lagging indicator that actually matters is still revenue.

With that said — here's what we ship.

1. Entity consolidation

Answer engines reason about entities — brands, people, products, places — and the relationships between them. If your brand exists in eight slightly-different forms across the web ("Acme," "Acme Co.," "Acme Inc.," "Acme Studios LLC"), the engines hedge. They cite the version they're most confident is real, which is often the wrong one.

Consolidate. Pick one canonical brand name. Update every directory listing, every social profile, every Wikipedia mention, every Crunchbase entry. Schema-mark every page on your own site with the canonical name. This is unglamorous and it works.

2. Schema and structured data

Add JSON-LD schema for: Organization, Person (for founders and named authors), Article (on every long-form piece), Product (if you sell things), and FAQPage (where appropriate).

Use sameAs liberally — it's the property that links your entity across LinkedIn, Wikipedia, Wikidata, Crunchbase, your social profiles, and elsewhere. Engines trust entities they can triangulate.

3. Author-attributed content

Anonymous blog content doesn't earn citations the way it used to. Every article we publish for clients now goes out under a named author with a credible biography, a real LinkedIn, and a body of work the engines can find.

That author should have topical consistency. An SEO column from a CFO is less useful, in citation terms, than the same column from someone whose entire LinkedIn is search-engine work. Be selective about who you put a name under.

Engines trust entities they can triangulate. Triangulation requires consistency you've earned, not consistency you've claimed.

4. Citation-worthy primary research

If you publish a piece of original data — a small survey, a benchmark study, a methodology — and other sites cite you when discussing the topic, the engines pick that up. Aggregator content gets summarized. Cited content gets named.

This is the single highest-leverage thing most brands aren't doing. A 200-respondent survey, well-designed and well-written-up, will out-rank 50 think-pieces — and get pulled into AI answers as the canonical source.

5. Cross-platform mention density

Engines reward presence across the open web. Not links, exactly — mentions. Podcast appearances. Quoted in trade publications. Speaking at conferences. Cited in industry reports. Mentioned on Reddit threads. The density of these signals — across diverse sources — is what tells the engine you're a real, relevant entity in this category.

This is slow, expensive, and unfakeable. Which is why it works.

6. Wikipedia and Wikidata presence

For brands above a certain notability threshold, a Wikipedia page is the single biggest unlock. ChatGPT and Perplexity both lean heavily on Wikipedia for entity grounding. Getting (and keeping) a page requires meeting Wikipedia's notability standards — independent reliable sourcing — and following its editorial norms. This is real work.

For brands below that threshold, focus on Wikidata, which is more permissive and feeds the same downstream systems.

7. Direct partnerships and submissions

Some answer engines accept direct content submissions and partnerships. Perplexity has a publisher program. OpenAI has data-partnership conversations with publishers and brands. These are early, occasional, and worth tracking — but not yet a reliable channel.

What we don't do

We don't:

• Pay for "AI SEO" tools that promise to optimize for LLMs. The space has more vendors than substance right now.
• Stuff pages with FAQ-formatted content unless the FAQ is genuinely useful to a reader. Engines see through this; readers leave.
• Chase low-quality citations from sites the engines don't trust. A mention in a content-farm aggregator doesn't help.
• Optimize obsessively for one engine. ChatGPT, Perplexity, Claude, and Gemini cite differently. Doing the fundamentals well is more durable than gaming any single system.

What to track

We track three things for clients each quarter:

Citation frequency. How often does your brand appear in answers to ten or twenty representative category queries? We run these manually each quarter; tools exist but their reliability is inconsistent.

Citation quality. When you appear, are you the primary source or one of five? Are you described accurately? Are you cited for the right thing?

Pipeline signal. When new prospects arrive, do any of them mention having found you through an AI answer? Right now this number is small but growing. It's the leading indicator that will eventually matter most.

That's the working playbook. It will change. We'll update this piece in May.