How to get cited by ChatGPT in 2026.

The behavior change is already measurable. When a buyer asks "who are the best Klaviyo consultants in Miami" or "is AdMax a real agency", the first surface they see is no longer a list of blue links. It is a synthesized paragraph, in ChatGPT or Google AI Overviews or Perplexity, with three to seven citations. Most buyers read the paragraph and either click one citation or do not click at all.

The implication for a small business is direct: if your brand is not in those citations, you are invisible in the moment the buyer is forming an opinion. Even when you rank #1 in traditional Google for the same query, the AI-generated answer above the blue links may not mention you. That gap — between traditional rank and AI citation — is the AEO+GEO surface. We close it with structural fixes to the website plus deliberate off-site placement.

Every model has a slightly different system, but the citation pattern across ChatGPT, Claude, Perplexity, Google AI Overviews, and Gemini converges on five signals:

1. Answer-shaped content — pages structured as Q→A blocks, definitions, comparisons, or how-to steps that match the conversational query.
2. Clean schema — Organization, Article, FAQPage, HowTo, and Service schema blocks that tell the model what the page is about without ambiguity.
3. Knowledge graph trust — the entity (your brand) has consistent descriptions across Wikipedia, Crunchbase, LinkedIn, and other authoritative knowledge graphs.
4. llms.txt and ai.txt declarations — explicit signal files that tell AI crawlers what to read first and how to summarize your site.
5. Off-site mention density — your brand appearing inside other sites the model already trusts (trade press, podcast transcripts, public GitHub READMEs, Reddit threads, association directories).

Models do not weight these signals equally and they do not publish the weights. But across every audit we have run, brands that move all five lines from red to green get cited within 90 days. Brands that move two or three get cited intermittently. Brands that move none do not get cited at all.

These are the structural changes we ship for every AdMax AEO+GEO engagement, ordered by impact-per-hour. You can do them yourself or have us do them via the $25 audit and follow-on retainer.

1. Publish an llms.txt at the root. A short Markdown file that tells AI crawlers what your site is, who you serve, and which pages summarize the brand best. Five minutes of work, immediate effect on Claude and Anthropic-adjacent crawlers.
2. Add Organization + WebSite schema to your root layout. JSON-LD blocks with a stable @id, your description, your founding date, your services, and your sameAs links to authoritative profiles (LinkedIn, Crunchbase). This is how every model knows what your brand is.
3. Build cornerstone pages with H1/H2 question shapes. Every cornerstone (services, FAQ, about, glossary) becomes a chance to answer a buyer question. The H2 is the question, the first paragraph is the direct answer in 40 to 80 words, the rest of the section is the supporting context. Models extract the direct answer; humans read the context.
4. Cover the comparison searches."X vs Y", "X alternatives", "is X worth it" — these are high-intent queries with disproportionate citation rates. Write the comparison page even if you are not the preferred answer; honest comparisons get cited more than spammy ones.
5. Earn 5 to 10 off-site mentions in 90 days. Trade press, podcast appearances, an open-source repo with a README, an explainer thread on a niche subreddit. The mentions do not need to be in the New York Times — they need to be in places the model already trusts.
6. Track mentions in a tool, not in your gut. Search Atlas, Profound, or Otterly will tell you when and how each model cites you. Without measurement you cannot tell which off-site move worked.
7. Iterate the citations you do get. When ChatGPT cites you for one query, study what made it cite you, then build the same shape for adjacent queries. The first citation usually unlocks five more in the same topic cluster.

Weeks 1 to 2 — the structural moves ship. llms.txt, schema, cornerstone H2 reshaping, the first off-site placement attempts go out. Mention tracking baseline is captured.

Weeks 3 to 6 — the first citations appear, almost always in Perplexity first because it recrawls the fastest. Claude and ChatGPT lag by one to three weeks behind Perplexity. Google AI Overviews tend to lag the longest because they piggyback on the regular Google index.

Weeks 7 to 12 — the citation count compounds. By day 90 every client we have run this play for has at least 10 citations across the four major engines for queries in their category. Some have over 100. The variance is driven by category competitiveness, not by execution quality.

The playbook above is honest. You can run all seven moves yourself if you have an in-house marketer who can write schema, draft cornerstone pages, pitch podcasts, and set up a tracking tool. Plan on 80 to 120 hours of focused work over the first quarter, then 5 to 10 hours a week of upkeep.

The case to hire AdMax is the obvious one: senior strategist plus AAAT (four AI agents that handle iteration, optimization, drafting, and competitive watch) ship the same play in weeks, not quarters, and the citations land faster. The $25 audit measures where you are starting from. The retainer starts at $300 per month and is cancellable any time.