When someone asks ChatGPT “which agency do you recommend for redesigning my website?”, that answer comes from somewhere. For the last twenty years, the battle for visibility was fought on Google and on social media; now it’s moving toward conversational assistants, and more and more users ask them directly which brand to choose. It’s no longer enough to appear in Google’s top results — the question now is what ChatGPT says about your brand, and whether it recommends you.

What GEO is and why it matters

GEO (Generative Engine Optimization) is the set of strategies for having your content understood, selected, cited and recommended by engines based on generative AI. While traditional SEO seeks to rank a page in search results, GEO seeks to make your brand part of the answer the AI gives directly to the user — in a world where more and more searches are resolved without a single click, being cited within that answer is the new indicator of digital authority.

Gartner projects that traditional searches could fall by up to 25% as AI assistants grow. Brands that don’t work on their presence in these environments risk becoming invisible to a growing portion of their potential customers.

How generative engines think and decide

Generative engines combine language models with real-time information retrieval systems: given a question, they analyze their corpus of pages looking for relevant content and synthesize an answer from the sources they consider most reliable. They don’t search like a traditional search engine — they learn from millions of texts and prioritize sources with verifiable data, narrative consistency across platforms and authority in a specific territory. Generic content, however much volume it has, doesn’t get cited.

Research from Princeton and Georgia Tech found that optimizing content for generative engines can increase a source’s visibility in their answers by up to 40%. That gap between brands that work on their presence in LLMs and those that don’t is going to keep growing.

Strategies to get ChatGPT to recommend you

  • Answer specific queries clearly. Research which concrete questions users ask about your industry (tools like AnswerThePublic or AlsoAsked help find them) and answer them in the first paragraph of the content, before qualifying or expanding. AI prefers direct, concise answers.
  • Create original content with your own data. The web is filling up with AI-generated content, and models detect that pattern. Your own data, methodologies, verifiable comparisons or field studies make you a source that other sites have to cite in order to reproduce that information.
  • Structure content with AI in mind. Clear, hierarchical headings (H2, H3), short paragraphs centered on a single idea, lists and tables where appropriate, key conclusions in the first paragraphs, and well-structured frequently asked questions.
  • Build authority through third parties. Unlike SEO, in GEO most brand mentions in AI answers originate on third-party sites, not on your own domain: industry rankings, “best of” lists, reviews, press coverage and client case studies. 74% of the references that ChatGPT and Gemini treat as authoritative sources come from news media — press presence directly feeds the knowledge AIs work with.
  • Maintain entity consistency across the web. Your company’s name, brands, executives, locations and products must match across your site, professional profiles, directories, marketplaces and press releases. Contradictions make the AI confuse companies or use outdated information.
  • Update content periodically. Generative engines value the currency of information and penalize outdated material — include visible dates and recent data.
  • Allow AI bots access. For ChatGPT to discover, summarize and cite your content, your robots.txt has to explicitly allow the OAI-SearchBot crawler (different from GPTBot, which is used for training — you can block one and allow the other according to your editorial policy).
  • Check your Bing rankings. ChatGPT uses Bing results as the basis for a good part of its answers. If your content isn’t well indexed there (check it with Bing Webmaster Tools), it’s much harder for it to be cited.

The new indicator: Share of Model

Share of Model measures how often a brand appears in AI-generated answers compared with its competition — an indicator that is starting to complement traditional metrics such as share of voice in media or search engine rankings. The question is no longer just appearing on Google or Instagram, but being one of the answers the AI offers directly when someone asks for a recommendation.

GEO doesn’t replace SEO, it complements it

AI engines still feed on traditional content, and what’s already well worked for SEO is more efficient and indexable than what isn’t. The relationship is symbiotic: good SEO is the foundation for achieving visibility in AI tools too. Loading speed, structure, security and quality backlinks remain relevant, because generative engines value them too.

Generative engines don’t “browse” your site like a user. They process it as a source of citable facts.

The good news is that the technical foundation is the same one you should already have for SEO. GEO doesn’t replace it: it extends it one step beyond the traditional search result.