What is generative engine optimization (GEO)? The definition, GEO vs SEO, and what to change on your site

The short answer

Generative engine optimization (GEO) is the work of making your website's pages the sources that AI answer engines, such as ChatGPT search, Perplexity, Microsoft Copilot and Google's AI Overviews and AI Mode, retrieve, quote and link when they write an answer. A search engine ranks pages in a list; a generative engine reads a handful of pages and writes one answer from them, so the goal moves from "rank this page" to "be one of the sources the answer is built on".

In practice, GEO is SEO plus three things: letting the AI search crawlers in, writing facts a machine can quote with confidence, and being mentioned on the sites those engines already trust. This page is the definition. For the step-by-step, platform by platform, read how to appear in AI answers.

Where the term comes from

The name comes from a research paper, "GEO: Generative Engine Optimization", by Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan and Ameet Deshpande, researchers at Princeton University and IIT Delhi. It was first posted in November 2023 and published at KDD 2024, the ACM's data mining conference in Barcelona.

The authors built GEO-bench, 10,000 queries from 25 domains, and a test engine that answered each query from the top five Google results using GPT-3.5. They then rewrote one source page per query in nine different ways and measured how much of the answer that page earned. What they found:

• Adding citations, quotations and statistics helped most. The best methods improved visibility by up to 40% over the unedited page on their main measure.
• Keyword stuffing didn't work. Repeating the query's words, the classic SEO trick, performed worse than the baseline.
• Results varied by subject. What helped a history question didn't always help a law or science question.
• The effect held on a real engine. On Perplexity they measured improvements of up to 37%.

Treat those numbers as a lab result: a 2023 model answering from five Google results is not today's ChatGPT or AI Mode. The direction has held up, though: pages that state checkable facts, with their sources, are easier for an answer engine to use.

GEO vs SEO

GEO doesn't replace SEO; it adds a second place to win. The table shows what changes.

The overlap is large because most answer engines start from a search index. Google says its AI Overviews and AI Mode use retrieval from its core Search ranking systems and "query fan-out", a set of related searches the model runs in parallel to gather more pages. A page that doesn't rank for any of those searches is unlikely to be read.

What Google says about GEO

Google answers the question directly in its guide to generative AI search, last updated on 10 July 2026: "From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO." The guide adds two eligibility rules for Google's AI features:

• The page must be indexed and eligible to be shown in Google Search with a snippet.
• The site must be included in Search generative AI features in Search Console. Inclusion is the default; the control, rolled out to all sites by 31 August 2026, lets an owner opt out.

The same guide lists "hacks" Google Search ignores: llms.txt files and other special markup, "chunking" content into tiny pieces, rewriting pages just for AI, seeking inauthentic mentions, and special structured data. It also warns that creating separate pages for every variation of a question, to manipulate AI responses, "violates Google's scaled content abuse spam policy".

Other engines document less. OpenAI's crawler page is the clearest statement outside Google: "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links."

What GEO adds to SEO: eight site-level actions

These are the changes that apply to almost any business site. Each one helps people too, which is why it also survives Google's rules.

1. Let the AI search crawlers in. Allow OAI-SearchBot, ChatGPT-User, PerplexityBot, Claude-SearchBot and Googlebot in robots.txt, and check that your firewall or CDN isn't blocking them anyway. You can allow search crawlers while blocking training crawlers such as GPTBot; OpenAI treats the two settings independently. Our AI visibility checker tests each crawler against your robots.txt.
2. Put your facts in plain HTML. Prices, plans, service area, team size, process, policies: each on a page with its own address, in text the server sends, not text that appears only after JavaScript runs. Dardo publishes its own on a facts page.
3. Answer first. Open each page and each section with the answer in one to three sentences, with the number, the year and the name in it, then explain.
4. Show your sources. The paper's strongest methods were citations, quotations and statistics. Link the primary source for each claim, and quote your own measurements with their dates.
5. Write one page per real question. A clear page that matches the question beats ten near-duplicates, and Google treats mass-produced variants as spam.
6. Say who you are in the quotable sentences. "Dardo is a team of three in Bogotá" survives being lifted out of context; "we are a small team" doesn't.
7. Earn genuine mentions. Answer engines read directories, reviews, roundups and forums. Correct what they say about you, and get listed where buyers in your category already look. Buying mentions doesn't help: Google's guide calls inauthentic mentions out by name.
8. Keep structured data for what it does. It makes pages eligible for rich results in Google; there is no special schema for AI answers. An llms.txt file is optional; see what llms.txt is and who reads it.

How to measure GEO

Answers change from one day to the next, so measure trends, not single answers.

• Search Console's Generative AI performance report shows impressions from Google's AI features for your site.
• Bing Webmaster Tools' AI Performance report, in public preview since February 2026, shows how often Copilot and Bing's AI answers cite your pages.
• A fixed prompt panel: 30 to 50 questions your buyers ask, run on the same engines once a month, counting whether your brand is named and whether your site is cited, separately.
• Your contact form: ask "How did you hear about us?" with an AI assistant as an option. Many AI visits arrive without a referrer, so analytics alone undercounts them.

Be careful with any tool that promises a single "AI visibility score". Google's guide says it plainly: "No third-party tool has access to our internal ranking or AI systems."

What GEO can't promise

Nobody controls what an answer engine writes. Engines change models and sources without notice, the same question can return different brands on different days, and an unknown brand with no mentions elsewhere rarely appears however well its own pages are written. GEO raises the odds that your pages are read and quoted; it doesn't guarantee it.

How Dardo works on GEO

Dardo's generative engine optimization service starts with an AI visibility audit, from USD 690 one-time: crawler access, how your facts read to a machine, what answer engines say about you today and where they get it. Monthly tracking with a prompt panel and reporting starts at USD 299 a month. The technical side of GEO overlaps with SEO, and most fixes land in the same pull request.