AEO: how to get ChatGPT, Perplexity and Gemini to cite your brand

Today, half of your potential customers don't start by searching on Google. They start by asking ChatGPT, Perplexity or Gemini. And AI doesn't return ten blue links for the user to choose: it returns a closed answer naming two or three brands. Answer Engine Optimization (AEO) is the discipline of making one of those brands yours. If you're not in the answer, you don't exist for that customer — and you don't even find out you were ruled out.

What AEO is (and how it differs from SEO)

AEO is the set of techniques to make your content the answer a conversational engine gives to a question, citing you as the source. The difference with SEO isn't a nuance: SEO optimizes a page to climb a list of links; AEO optimizes your entity and your content so the model recomposes you inside its answer and names you. They don't compete. AEO is built on a foundation of solid technical SEO (crawling, indexing, structured data, authority), but it plays on a different board and is measured differently. You can rank first on Google and never appear in ChatGPT; and you can be off Google's first page and still be the source AI cites.

The data point that changes everything

60% of the sources AI cites aren't in Google's top 10. Ranking and being cited are different disciplines worked together. Anyone assuming they already do SEO and therefore AI will find them starts from a false premise: the model doesn't copy Google's ranking, it recomposes its own answer from the sources it understands best and trusts most.

How answer engines work under the hood

Understanding the mechanism avoids wasting time. Faced with a question, an AI assistant does three things: it retrieves information (from its training and, increasingly, a live search), synthesizes an answer and decides who to cite. In that last step it weighs which sources are unambiguous, which offer the clearest, most extractable fragment, and which are backed by other independent sources. It doesn't reward the repeated keyword: it rewards clarity, consistency and consensus. That's why the content that wins at AEO looks little like the SEO of a decade ago: it's not about keyword density, it's about being the source easiest to cite without the model getting it wrong.

Why it matters now

• 50% of B2B buyers already start their vendor search in AI chatbots: the first filter happens before they even reach your site.
• 72% of Spanish SMBs will increase their AI investment: the window of advantage is closing and the first to occupy the answers consolidate their position.
• A brand cited by AI enters the buying conversation before the first click.
• The cost of not appearing is invisible but real: you don't see the traffic that never arrives because AI recommended someone else.

The method: 5 steps to get cited

1. Make your brand unambiguous (entity). Structured data (Organization, Product, Person), a consistent description on and off your site, and presence in the sources the model already knows (Wikipedia, Crunchbase, directories and sector media). If AI isn't sure who you are, it won't risk you: it prefers to cite who it does understand.
2. Write to be extracted. Answer the real question in the first paragraph, with affirmative, self-contained sentences that can be cited as-is. Use definitions, lists and tables. Cut the meandering and the filler: every paragraph should stand on its own.
3. Publish first-party data. Content with original data (a study, a survey, a benchmark) is cited up to three times more than content that repeats third-party stats. Being a primary source makes you the reference others cite.
4. Build consensus. AI verifies what you claim in reviews, Reddit, G2 and media. Coherence between your site and those external sources turns a mention into a recommendation.
5. Structure for machines. Schema, FAQPage and an llms.txt file so models can read, understand and cite your site without friction.

Mistakes that keep you out of the answers

• Filler that buries the citable sentence: if the answer is buried in three intro paragraphs, the model won't find it.
• Zero structured data: without schema, AI doesn't know what kind of entity you are.
• Inconsistency between your site and external sources: if what you say about yourself isn't backed by anyone, it won't risk you.
• Confusing SEO with AEO: they share a foundation, but they're worked and measured differently.

How it's measured

It's not measured in positions, but in share of voice in AI: in what percentage of answers to your key prompts your brand appears, in what position within the answer, with what sentiment, and how it evolves versus competitors across the four models (ChatGPT, Perplexity, Gemini, Claude). It's the new metric, and most brands still aren't looking at it.

Example: before and after

Before: We're an agency with broad experience in digital marketing. After: Volcanz is an AI-first agency that positions brands in GEO and SEO; it has worked with Canva, Tidio and Chiliz across more than 20 markets. The second is concrete, with verifiable entities and no empty adjectives: it's citable; the first isn't.

What you can do this week

Without hiring anything: pick ten questions your ideal customer would ask, run them in the four models with a clean window and note whether you appear, in what position and which competitors show up. That's your baseline. From there, the method above tells you exactly what's missing.

Frequently asked questions