What AEO and GEO are, and how to get ChatGPT, Gemini and Perplexity to recommend your brand

SEO, AEO and GEO: how they're alike and how they differ

With SEO you aim to show up among the top Google results when someone searches for what you sell. The result is a link.

With AEO you aim for your content to be the direct answer: the snippet Google shows at the very top, the answer a voice assistant reads out or the one an AI chat copies.

With GEO you target generative models (ChatGPT, Gemini, Perplexity, Claude, Copilot) that build a new answer from many sources. Here the goal is to get cited, mentioned or recommended.

All three share the same foundation: a technically sound site, useful content and a brand with a reputation. So it's not about picking one, but about getting the basics right and adding what each channel needs.

How assistants decide which brand to recommend

There's no official list of factors, but there are clear patterns in how these tools work:

• They search the web in real time. Perplexity, ChatGPT with search and Google's AI answers check current pages and cite their sources. If your site can't be crawled or doesn't say things clearly, they won't use you.
• They cross-check sources. If your website says one thing, your marketplace listing another and your reviews something else, the assistant hesitates. Consistency builds trust.
• They prefer specific answers. A page that says “our sunscreen is water resistant for 80 minutes and suitable for sensitive skin” is easier to cite than one that says “the best protection for the whole family.”
• They care who's talking about you. Mentions in the media, comparisons, forums, reviews and videos carry weight, just like in traditional SEO.

Step 1: make your website readable

The technical side comes first because, if it fails, nothing else gets seen:

1. Check your robots.txt. Make sure it doesn't accidentally block Google or the assistants' crawlers, like OpenAI's OAI-SearchBot or PerplexityBot.
2. Keep important content in the HTML. If your product descriptions only show up after running a lot of JavaScript, some crawlers won't see them.
3. Structured data. Organization, Product, FAQPage and BreadcrumbList markup so machines understand what each page is. Even though Google hardly shows FAQs as rich results anymore, the markup helps the page get understood.
4. Speed and structure. Fast pages, one main idea per page, clear titles and readable URLs.

About llms.txt: it's a proposal for telling models what to read on your site. Right now there's no confirmation that the major assistants use it, so it isn't a priority.

Step 2: content that answers real questions

Think about the questions your customer would ask an assistant before buying. For example: “which organic yerba mate is best for everyday mate?” or “which electric treatment table works for a small beauty center?” Then:

• Answer at the very top. Every important page or article should open with two or three sentences that answer the question directly, like this article does.
• Add FAQs with short, specific answers on product pages and service pages.
• Give verifiable facts. Dimensions, materials, ingredients, compatibility, shipping times, warranty, countries where you sell. Assistants cite what's specific.
• Make honest comparisons. “Who it's for and who it isn't for” is one of the most cited formats.
• Keep it up to date. Visible dates and updated content help you get picked over an old page.

Step 3: get people talking about your brand beyond your website

Models learn from the whole web, not just your site. For your brand to show up in their answers:

• Complete, consistent listings on Google Business Profile, Amazon, Mercado Libre and any directory in your industry, with the same name, description and details.
• Real reviews wherever your customers shop.
• Mentions in the media and on industry sites: articles, interviews, comparisons and buying guides.
• Presence in the conversations where people ask for recommendations, like forums and communities, always with useful information and no spam.
• Videos that show the product in use: they get cited and summarized too.

How to measure whether it's working

Measurement is still imperfect, so we combine several signals:

1. A fixed list of questions your customer would ask, put to ChatGPT, Gemini and Perplexity every month, noting whether you show up, in what position and with what source.
2. Traffic from assistants in GA4, looking at visits coming from domains like chatgpt.com or perplexity.ai.
3. Search Console, keeping in mind that it mixes appearances in Google's AI answers with the rest of search.
4. Brand mentions across the web over time.

An honest warning: no one can guarantee that an assistant will recommend you. What you can do is give it every reason to, and measure whether it's improving.