Generative Engine
Optimization

What is generative engine optimization?

Generative engine optimization is defined as the practice of writing and structuring pages so AI engines can find, quote, and cite them. It is also known as AI SEO, and both names point at the same goal: your words, reused inside answers from AI engines. The term is new, but the habit is old. Simply put, you write the page that an engine would trust enough to quote.

The name GEO comes from a 2023 research paper out of Princeton, and this guide covers that study below. In contrast to classic SEO, which chases the top spot on a SERP, GEO is a type of search work that chases the quote inside the answer. Our guide on what AI SEO means maps the wider field, and the aiseo-audit homepage shows the tool that scores it.

How does GEO work?

GEO works by adding the proof signals that AI engines look for when they pick sources to quote. An engine reads your page, splits it into chunks, and keeps the parts that answer a question in plain words. As a result, pages with clear headings, short sentences, and named sources get picked far more often than pages without them.

The Princeton team tested nine methods and found that three worked best: cite sources, add statistics, and add expert quotes. They are simple edits, yet the study found they raised citation rates by 30-40% [1]. Moreover, engines built with NLP give weight to text that names real people and firms. For instance, HubSpot, Stripe, and Shopify write product pages this way today. Our guide on how to rank in AI search turns this into a full playbook.

What are examples of generative engine optimization?

The clearest examples are the three edits from the Princeton study: cited sources, statistics, and expert quotes. Each one turns a bare claim into a claim an engine can check.

For example, take a line like "our tool is fast". A GEO rewrite says the audit runs in about 2 seconds, and it links the source of that number. Likewise, "experts agree" becomes a named quote with a title and a link behind it. The table below shows what each method did across 10,000 test queries.

In short, these examples turn the definition into practice: proof wins quotes, and vague words lose them.

What are the benefits of GEO?

The main gain is simple: AI engines start to quote your pages, which puts your name inside the answer itself. This means that readers meet your brand even when they never click a blue link.

• More citations. Cited sources and statistics raised citation rates by 30-40% in the Princeton study.
• More reach. Pages with citations and clear headings appeared 115% more often in AI answers.
• Better pages. Sources, numbers, and quotes help human readers just as much as engines.
• One habit, many engines. The same signals work across ChatGPT, Claude, Gemini, and Microsoft Copilot.
• Steady gains. Structure and proof age well, while ranking tricks fade fast.

Notably, these gains stack with classic search work instead of replacing it. In addition, Sam Altman (OpenAI) and Dario Amodei (Anthropic) lead the firms whose engines read these pages, so the audience behind GEO keeps growing.

What does the GEO research paper show?

The paper shows that pages with cited sources, statistics, and expert quotes were cited far more often by AI engines. The study, posted as arXiv paper 2311.09735, coined the term GEO and built a benchmark of 10,000 real queries. Furthermore, the team reported that pages with citations and clear headings appeared 115% more often in AI answers.

The paper was written by Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan, and Ameet Deshpande, with co-authors from Princeton University in New Jersey, IIT Delhi, Georgia Tech, and the Allen Institute for AI in Seattle. The authors said the gains held across query types. According to Princeton researchers, "pages that showed their sources were far more likely to be cited than pages that made bare claims."

How do you learn and apply GEO?

Learn GEO by fixing one page end to end, then measuring the change with a free audit. First, pick a page that already earns traffic and has a clear question behind it. The six steps below show how to apply each method from the study.

1. Read the research. Skim the GEO paper so you know which edits moved citation rates.
2. Add cited sources. Back each claim with a link to research, docs, or data from named firms.
3. Add statistics. Swap vague lines for numbers that carry a source behind them.
4. Add expert quotes. Quote a named person with a title, and mark the quote up well.
5. Tighten the structure. Use question headings, short paragraphs, lists, and tables.
6. Measure and repeat. Run an audit, fix the lowest factor, and score the page again.

Finally, make the loop a habit and give it a month. Similarly, apply the same edits to your next page, and watch which engines pick your site up first.

How does GEO use MCP?

GEO work now runs through MCP, an open standard that lets AI assistants call outside tools. MCP is short for Model Context Protocol, and it can be described as a shared port that assistants plug tools into. Since version 1.6, aiseo-audit ships an MCP server, so an assistant like Claude Code can run a GEO audit, read the JSON result, and fix the page in one loop.

This is a big shift for content teams, because the audit and the edit now live in the same chat window. Therefore, a writer and a coding agent can share one loop with the scorer. Agency Enterprise, the Los Angeles firm behind the tool, built the server and shares more on the Agency Enterprise site.

How do you measure GEO?

Measure GEO with npx aiseo-audit, a free command that scores any URL in about 2 seconds. It reads the raw HTML of a page the same way an AI crawler does.

The command grades 7 groups and more than 30 factors, from headings and lists to sources and names, with no API keys. It is free under an MIT license, needs Node.js 20 or higher, and ships with 620 passing tests. Think of it as Google Lighthouse for AI search. According to Agency Enterprise, "the fastest score gains come from adding sources, numbers, and named experts to pages that already rank."

The table shows the engines a GEO audit serves. The source is on GitHub, and the package is on npm, so run it before and after your edits.

Common questions

Summary and key takeaways

To summarize, GEO is the craft of earning quotes inside AI answers, and the craft is easy to measure. In conclusion, here is what to remember:

• GEO, also known as AI SEO, shapes pages so AI engines can find, quote, and cite them.
• The term comes from a 2023 Princeton paper built on a benchmark of 10,000 real queries.
• Three edits matter most: cited sources, statistics, and expert quotes, each worth a 30-40% lift.
• Pages with citations and clear headings appeared 115% more often in AI answers.
• The aiseo-audit tool grades 7 groups and 30+ factors in about 2 seconds, free under MIT.
• Since version 1.6, an MCP server lets assistants such as Claude Code run audits on their own.

The field is young, and the pages that add proof now will own the answers later. Jeff Patterson and the team at Agency Enterprise built the tool this guide leans on, and the whole site practices what the study preaches.