What is generative engine optimization (GEO) and how to get into ChatGPT, Gemini and Perplexity answers

What GEO means in simple terms

More and more people no longer scroll through ten blue links. They ask an AI assistant: “Which dentist in Tallinn should I choose?”, “Where can I order a cake for a birthday?”, “Which CRM is best for a small company?” — and get a ready answer with two or three names.

GEO is everything a business does so that its name is among those two or three. The same thing is also called AI search optimization, AI SEO, answer engine optimization (AEO) or LLM optimization (LLMO). The names differ, the goal is the same: visibility in AI-generated answers.

Where the term came from

The term was introduced in the research paper GEO: Generative Engine Optimization by Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan and Ameet Deshpande (Princeton University, IIT Delhi and independent researchers). It was published on arXiv in November 2023 and presented at the KDD 2024 conference. According to the authors, GEO methods can increase a source’s visibility in generative engine responses by up to 40%: adding quotations, statistics and citations of sources worked best, while keyword stuffing gave little to no improvement.

How AI decides whom to recommend

An AI assistant does not “know” your company the way a person does. When it answers a question, it usually does two things:

1. Finds sources. ChatGPT Search uses third-party search providers, one of which is Bing; Copilot sends generated search queries to Bing; Gemini can ground its answers in Google Search; Google AI Overviews show pages from the Google index. If your website is poorly indexed, the AI may simply not find it.
2. Builds the answer. From the sources it found, the model takes the passages that answer the question best and names the companies mentioned in them.

What helps, according to the GEO research and our practice:

• Quotes, numbers and sources. In the GEO paper, adding quotations, statistics and citations increased visibility the most; keyword stuffing did not help.
• Consistent facts. In our practice, when the website says one price and a directory another, AI answers repeat the mistake or skip the company.
• Structured data. Schema.org markup describes the company, services and prices in machine-readable form. Google states that no special markup is required for AI Overviews, but structured data must match the visible text on the page.
• Clear answers. Pages that answer a specific question in the first two or three sentences are easier to quote.
• Independent mentions. Directories, media, industry sites and reviews confirm that the business exists and is worth recommending.
• Indexing. For Google AI Overviews and AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet; there are no other technical requirements.

GEO vs SEO: what is the difference

GEO does not replace SEO. AI assistants find sources through search indexes, so a site that is invisible in Google and Bing is usually invisible to AI too. That is why at JW2D we run SEO and GEO as one system.

Is SEO dead because of AI?

No. “Is SEO dead now with AI?” is one of the questions Google shows in “People also ask” for GEO searches (AnswerThePublic data, October 2026), and the honest answer is that SEO is changing, not disappearing. AI answers are built on indexed pages: Google, for example, only shows pages that are indexed and eligible for a snippet as supporting links in AI Overviews. Good SEO is now the entry ticket, and GEO decides whether you are actually named.

How to do GEO: step by step

1. Audit. Collect 20–50 real questions your customers ask and check what ChatGPT, Gemini, Perplexity, Claude and Copilot answer today. Are you mentioned? Who is mentioned instead? Which sources are cited?
2. Put the facts in order. One version of the company name, address, phone, prices and services everywhere: website, Google Business Profile, directories, social profiles.
3. Add structured data. Schema.org markup for the organization, services, prices, FAQ and articles.
4. Write answers, not slogans. Pages and articles that answer specific customer questions in the first sentences, with numbers, quotations and links to sources.
5. Grow independent mentions. Directories, industry media, partner sites and real customer reviews.
6. Monitor. Repeat the same questions regularly and track changes.

Avoid fake reviews and bought mentions. Models learn to devalue them quickly, and for a brand it is a reputational risk.

How to measure GEO results

There is no “position 1” in an AI answer, so GEO is measured differently:

• Mention rate — in how many of your key questions the AI names your business.
• Position in the answer — first, second or at the end of the list.
• Accuracy — whether the AI describes your services, prices and location correctly.
• Sources — which pages and sites the model cites when it talks about you.
• Leads — how many customers say they found you through ChatGPT or another AI.

Is GEO worth it for a small local business?

Yes. In our practice, local and niche questions are where a small business can get into AI answers fastest, because few competitors work on them yet. For example, asked in Russian, Gemini names our client Prjanja first for “Where in Tallinn can I order designer gingerbread” and Smoked House second for “Where in Tallinn can I buy smoked fish”. Queries, screenshots and dates are on the AI case studies page.

How much GEO costs and how long it takes

At JW2D, AI search optimization starts from €1200. The final price depends on the niche, the number of markets and AI platforms, and the current state of the website and mentions. The audit and technical base are usually done in the first weeks; after that, the pace depends on how often each model updates its sources, so we show real progress in regular reports instead of promising a fixed date.