How to Show Up in ChatGPT, Perplexity, and Google AI Overviews: An Executive's Playbook
To show up in ChatGPT, Perplexity, and Google AI Overviews, you need three things the models look for: a clear entity they can identify, an owned site full of answer-first content they can read and cite, and third-party sources they already trust that name you. Most executives have none of these and wonder why the AI has never heard of them. Here is the playbook, in the order I actually run it.
Step 1: Test where you stand
Before anything, get a baseline. Open each assistant and ask two kinds of questions: a name query (“Who is [your name]?”) and a category query (“Who are the best people to help with [your specialty]?”). Note, for each engine, whether you are surfaced at all, described accurately, cited with a link, and actually recommended. That four-part read — surfaced, accurate, cited, recommended — is your scorecard. Most leaders find they pass the name test and fail the category test completely. That gap is the whole game.
Step 2: Become an unambiguous entity
AI assistants resolve people through structured, authoritative identity signals. Get these in place:
- Claim your Google Knowledge Panel (and correct any errors). This is the single strongest entity signal for Google AI Overviews and Gemini.
- Create a Wikidata entry with solid statements — education, role, employer, official site, and the identifiers (LinkedIn, Crunchbase) that tie your records together.
- Add
Personschema to your website so machines read your identity directly from a source you control. - Make your naming consistent everywhere — same name, same title, same one-line description — so the model doesn’t confuse you with someone else.
Step 3: Build the mothership
Stand up a website you own and fill it with answer-first articles — pieces that lead with the answer in the first sentence, then support it. This is what Perplexity and Google AI Overviews quote, because they extract clean, citable answers. Practical rules: one clear question per article, the answer up top, deep enough to be the best result, clean schema (Article, FAQPage), and a robots.txt / llms.txt that explicitly welcome AI crawlers. Publish transcripts of any talks or videos — the AI reads text, not footage.
Step 4: Earn third-party citations
This is the step that moves you from described to recommended. Models recommend names that appear on sites they already trust — so go get named there: guest on podcasts (each one is a citation and a transcript), write bylines for reputable outlets, get included in “best-of” and expert round-up lists, and seek earned press in your category. Perplexity in particular is citation-hungry and B2B-favored, so independent mentions there pay off fast.
Step 5: Measure, then repeat
Re-run the Step 1 panel on a schedule — biweekly works — and watch the trend. Discoverability is not a one-time project; it is a compounding practice, exactly like SEO was. The engines change constantly, so the work is staying current, not finishing.
The priority order, in one line
Entity first (Knowledge Panel, Wikidata, schema) → owned answer-first content → third-party citations → measure and repeat. Do them out of order and you will spin; do them in order and the models start naming you. The executives who run this playbook now, while their category is still uncrowded in AI answers, will hold those slots while everyone else is still posting into the void.