SEO + AIBy Mariano González Campas · 9 min read
What Is GEO: How to Get ChatGPT and Perplexity to Recommend Your Brand
GEO (Generative Engine Optimization) is optimization for generative engines: the set of practices that increase the odds that ChatGPT, Perplexity, Gemini, or Claude will mention, cite, or recommend your brand when someone asks about your industry. If SEO got you ranked in Google's list of results, GEO gets you inside the answer the AI writes.
More and more people are resolving their searches without ever touching Google: asking ChatGPT which agency to hire, asking Perplexity which product to buy, asking Gemini how to solve a problem. If your brand doesn't show up in those answers, you're invisible to a growing share of your market. In this guide we cover what GEO is, how it differs from SEO, and how to get started today, no hype.
Key takeaways
- GEO (Generative Engine Optimization) is the discipline of optimizing your digital presence so generative engines — ChatGPT, Perplexity, Gemini, Claude — cite and recommend your brand in their answers.
- GEO doesn't replace SEO: it extends it. They share a technical and content foundation, but GEO adds extractability, entity clarity, and presence in the sources AI consults.
- Models cite what they can understand, summarize, and attribute: clean definitions, direct answers to concrete questions, and data with context.
- The technical starting point is concrete: allow AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended), publish an llms.txt file, and mark up your site with schema.org.
- GEO is measured with canonical prompts: the questions your ideal customer would ask, checked every month across each engine to track whether you show up and who gets cited instead.
- Most Argentine companies still aren't doing GEO: starting now means claiming cheap real estate that's going to get expensive later.
What is GEO (Generative Engine Optimization)?
GEO is optimization for generative engines: working on your website, your content, and your digital reputation so AI models mention you when they write an answer. You'll also see it called AIO (AI Optimization) or AEO (Answer Engine Optimization) — the boundaries between the terms are still blurry, but the goal is the same: be part of the answer, not part of a list of links.
In practice, GEO combines three fronts. The technical one: making sure AI crawlers can read your site and find structured data. The content one: making your pages answer real questions in phrases a model can quote. And the authority one: getting other sources — media, directories, reviews — to talk about your brand, because AI inherits whatever its sources say.
Why GEO matters now
Search behavior is changing fast. Industry studies agree that a growing share of commercial searches — what to buy, who to hire, what tool to use — now starts directly in an AI chat or ends in a generated answer, like Google's AI Overviews. The user gets a written recommendation and often never even opens the links.
And here's the detail that changes the game: a generative answer usually names two, three, or four brands, not ten blue links. On Google, ranking fifth still gets you clicks; on ChatGPT, you're either among the brands named or you don't exist for that person. It's an all-or-nothing logic that rewards whoever moved first.
Does GEO replace SEO? What changes and what stays the same
No, GEO doesn't replace SEO: it extends it. Classic SEO competes for positions in a list of results; GEO competes to be part of a written answer. They share the same foundation — quality content, authority, structured data, a fast and crawlable site — which is why a site with solid SEO already has half the GEO work done.
- Stays the same: useful, original content, domain authority, clean architecture, speed, and structured data.
- The goal changes: ranking a URL is no longer enough — you want the model to understand your brand and name it.
- The winning format changes: self-contained paragraphs of 40 to 80 words get cited; sprawling wall-of-text articles don't.
- Measurement changes: from rankings and clicks to brand mentions per prompt, per engine, per month.
How do generative engines decide what to cite?
Each engine works differently — Perplexity searches the live web, ChatGPT combines its training with search, Gemini leans on Google's index — but the underlying criteria repeat. These are the five that carry the most weight:
- Sources they consult: engines with built-in search read web results, directories, reviews, Wikipedia, and media outlets. Being present in those sources is half the battle.
- Entity clarity: the AI needs to be able to answer "what is your brand?" in one sentence. If your value proposition is confusing, nobody cites you.
- Citable content: direct answers to concrete questions, clean definitions, data with context. Wall-of-text paragraphs don't get quoted.
- Structured data: schema.org (Organization, FAQPage, Service) gives models machine-readable, verifiable facts.
- Authority signals: mentions in media, reviews, links. AI inherits the biases of its sources — if nobody talks about you, neither will it.
How to show up in ChatGPT and Perplexity: a starter checklist
Good news: getting started with GEO doesn't require a corporate budget, it requires a method. Here's the checklist we use to kick off every project:
- Allow AI crawlers in your robots.txt: GPTBot (ChatGPT), ClaudeBot (Claude), PerplexityBot (Perplexity), and Google-Extended (Gemini).
- Publish an llms.txt with a summary of your business, services, coverage area, and key facts.
- Mark up your site with schema: Organization with sameAs links to your profiles, FAQPage with real customer questions, Service for each service you offer.
- Rewrite your homepage so the first sentence says what you do, for whom, and where. No "comprehensive solutions."
- Answer the 10 questions your customers ask most often on your site, in self-contained answers of 2 to 4 sentences.
- Build consistent mentions of your brand — same name, same description — across directories, media, and communities in your industry.
What is llms.txt?
llms.txt is a plain text file published at the root of your site (yourdomain.com/llms.txt) that summarizes, in simple format, what your business is, what it offers, and where to find the important information. It's an emerging standard — not every engine reads it yet, but implementing it takes an hour and gets you ready for the ones that do.
How is GEO measured?
You define a set of canonical prompts — the questions your ideal customer would ask, things like "what web design agency should I hire in Buenos Aires?" or "what e-commerce platform do you recommend for a small business?" — and check them every month across each engine: ChatGPT, Perplexity, Gemini, and Claude.
For each query, you log whether your brand shows up, in what position within the answer, with what description, and citing which source. That gives you a baseline you can track over time: more mentions, better descriptions, more of your own sources cited. It's still hands-on tracking, but it's the only serious way to know if you're gaining ground.
Common mistakes when starting with GEO
- Blocking AI crawlers "to protect your content": if you sell services or products, that's pulling yourself off the shelf with your own hands.
- Keyword-stuffing for the AI: models detect padded text better than Google does. Clarity pays off; filler doesn't.
- Abandoning SEO: generative engines feed on the same results SEO builds. No foundation, no GEO.
- Measuring only once: model answers change with every update. Without monthly tracking, you don't know if you're improving.
- Having a confusing entity: if your site says "we power digital synergies," no model will know what you sell or who to recommend you to.
Who should be investing in GEO today?
Any industry where the purchase decision involves research: professional services, software, healthcare, education, travel, high-ticket e-commerce. If your customers compare options before buying, some of them are already asking AI. And if you run a local business, those answers also cite local listings, reviews, and directories from your area.
The competitive edge is timing: in Argentina, the vast majority of companies still aren't working on GEO. The first ones to build entity clarity and citable content are going to be the reference the models keep repeating for years. After that, dislodging whoever's already being cited costs twice as much.
In SEO you compete for a ranking. In GEO you compete for a sentence in the answer.
Frequently asked questions
Does GEO replace SEO?
No. GEO extends SEO: they share quality content, authority, and structured data, but GEO adds extractability, entity clarity, and presence in the sources models consult. A serious strategy works both fronts at once: SEO gives you the crawlable foundation, and GEO turns it into mentions inside AI answers.
How long does GEO take to show results?
The technical changes in GEO — allowed crawlers, llms.txt, schema, citable content — can show up within weeks in engines that search the live web, like Perplexity. Brand authority and external mentions take months, same as in SEO. That's why monthly tracking with canonical prompts is part of the work, not an extra.
What is an llms.txt file?
llms.txt is a text file published at the root of your domain that summarizes, for AI models, what your business is, what services it offers, and where to find key information. It's a low-cost, emerging standard — it takes an hour to implement and improves how generative engines understand your site.
How do I know if ChatGPT already mentions my brand?
Ask it the way your customer would: "what web design agency do you recommend in Buenos Aires?" or "where's the best place to buy such-and-such product?" Repeat the same questions in ChatGPT, Perplexity, and Gemini, and note whether you show up and which sources get cited. That's your GEO baseline: homemade, imperfect, and miles better than measuring nothing at all.
Does GEO work for a local business in Argentina?
Yes, and significantly: when someone asks for a provider "near" or "in" a city, generative engines cite Google listings, reviews, and local directories. A local business with a clear entity, consistent reviews, and content that answers questions from its area has everything it needs to show up in those answers ahead of the competition.
Does blocking AI crawlers protect my content?
Blocking GPTBot, ClaudeBot, or PerplexityBot stops engines from reading your site, and with that, from citing or recommending you too. For a media outlet that makes money selling its content, that might make sense; for a business that wants customers, it's an own goal. It's a strategic call, but it's worth making it knowing exactly what you're giving up.
This very site you're reading applies everything above: llms.txt published, full schema markup, AI crawlers allowed, and citable content. We practice what we preach. If you want to know how generative engines see your brand today, at loco22 we start with a GEO audit and a free initial consultation: we show you where you show up, where you don't, and the concrete plan to change it.