ChatGPT now answers hundreds of millions of buying questions a week. Each answer names a handful of brands, and the selection isn't random — it follows patterns you can measure and influence. That practice has picked up a name: ChatGPT SEO.
The name is slightly misleading — there are no rankings to climb and no index to submit to — but the discipline is real. Here's how ChatGPT actually selects brands and sources, and the five-step playbook to become one of them.
How ChatGPT Selects Brands and Sources
Every ChatGPT answer draws on two systems:
- The model's prior — what it learned in training from listicles, reviews, forums, docs and news. This decides which brands "come to mind" for your category and how they're framed.
- Web search — for commercial and current prompts, ChatGPT searches before answering, decomposing the question into several hidden queries (query fan-out) and reading the results. Those retrieved sources heavily shape who gets named and cited.
Neither system reads your meta keywords. Both aggregate what the web says about you — which makes ChatGPT SEO closer to reputation engineering than to classic on-page work.
The 5-Step ChatGPT SEO Playbook
Open the door
Allow OAI-SearchBot and ChatGPT-User in robots.txt (GPTBot for training is your policy call). One legacy Disallow rule can erase you from answers — check in 10 seconds with the free crawlability checker.
Baseline your visibility
Track 30–50 real buyer prompts daily. Record mention rate, position, sentiment and — critically — which sources ChatGPT cites when answering.
Win the retrieved sources
The citation data gives you the exact listicles, review platforms and articles ChatGPT reads for your category. Getting into them moves answers within weeks — it's the highest-leverage work in the playbook.
Publish answer-shaped content
One page per recurring buyer question and fan-out sub-query: direct answer up top, question-form headings, FAQ schema. ChatGPT cites passages that map cleanly to what was asked.
Fix your entity
One consistent description and category label everywhere, Organization/Product schema, current third-party profiles. Models hedge on brands they can't pin down.
What Doesn't Work
- Keyword stuffing for the model. ChatGPT doesn't rank pages by keyword density; it synthesizes from sources it trusts. Optimizing prose "for the AI" instead of for the question wastes effort.
- Astroturfing communities. Reddit and forums matter enormously as signals, and both platforms and models are increasingly good at discounting fake advocacy. Founder-voice honesty compounds; sockpuppets get discounted.
- One-off checks as measurement. ChatGPT is non-deterministic — the same prompt returns different brands run to run. Decisions made on a single screenshot are decisions made on noise.
Measuring ChatGPT SEO
The KPI set mirrors classic SEO, translated to answers: mention rate (share of tracked prompts naming you — your visibility score), position (first recommendation vs afterthought), share of voice against competitors, citation rate of your pages, and downstream, AI referral traffic from chatgpt.com to your site.
Geonimo's ChatGPT Visibility Tracker automates the full loop — daily sampling, fan-out and citation capture, competitor benchmarks, plus attribution of ChatGPT-referred visits and leads on your site. The playbook above plus that feedback loop is, in practice, the whole discipline.
Start where every good program starts: a baseline. Get your free audit and see exactly how ChatGPT answers your buyers' questions today.
Related Articles
Frequently asked questions
What is ChatGPT SEO?
ChatGPT SEO is the practice of optimizing your brand and content so ChatGPT mentions, recommends and cites you in its answers. It targets both the model's training knowledge (what it learned about your category) and its web-search behavior (the sources it retrieves before answering).
Can you do SEO for ChatGPT?
Yes — not by manipulating the model, but by shaping the evidence it uses: presence in the listicles and review sites it retrieves, content that answers buyer questions directly, consistent entity signals, and crawler access for OAI-SearchBot and ChatGPT-User. All of it is measurable with daily tracking.
How does ChatGPT choose which websites to cite?
When it searches, ChatGPT fans the question into multiple queries, retrieves results, and cites pages that directly answer the question with clear structure and credible signals. Recommendation-style prompts favor third-party validators — comparison articles, review platforms and community threads — over vendor pages.
Does ChatGPT use Bing for search?
ChatGPT's search has historically drawn on Bing's index along with OpenAI's own crawling (OAI-SearchBot). Practically, that means classic search hygiene still matters — indexable, fast, well-structured pages — plus OpenAI-specific crawler access.
How do I check if ChatGPT recommends my brand?
Ask it your buyers' questions in fresh sessions and record who gets named — then repeat, because answers vary between runs. For a statistically meaningful picture, a tracker like Geonimo samples your prompt set daily and charts mention rate, position, sentiment and cited sources over time.
