GEO, AEO and SEO describe overlapping work: making sure AI search tools can find your pages, read them, understand who you are, and use you in an answer. Google says that, from its point of view, optimizing for generative AI search is still SEO. What changes is that ChatGPT, Claude and Perplexity send their own crawlers, and those crawlers don’t behave like Googlebot. Start with the conditions that come first: access and readable HTML.
The terms
| Term | Stands for | What people usually mean |
|---|---|---|
| GEO | Generative Engine Optimization | Getting cited or mentioned in generative AI answers. The term comes from a 2023 research paper |
| AEO | Answer Engine Optimization | Writing so an answer can be lifted directly from the page |
| AI search optimization | — | The umbrella term for both |
Short definitions are in the glossary.
What Google says
Google’s guide “Optimizing your website for generative AI features on Google Search” (updated 2026-07-10) says:
From Google Search’s perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO.
The same page says you don’t need AI-specific files, markup or Markdown to appear in Google’s AI features, that structured data isn’t required for them, and that you shouldn’t chunk or rewrite content just for AI. It also adds an eligibility condition beyond indexing: a site must be included in Search generative AI features in Search Console.
What is different outside Google
ChatGPT, Claude and Perplexity build their answers from pages fetched by their own crawlers — OAI-SearchBot, Claude-SearchBot and PerplexityBot. Two things that rarely matter for Google can stop them:
- robots.txt rules aimed at AI. Many sites block “AI bots” as a group and catch search crawlers along with training crawlers. OpenAI, for example, says sites that opt out of OAI-SearchBot won’t be shown in ChatGPT search answers. See AI crawlers and robots.txt.
- JavaScript. A 2024 measurement by Vercel found that none of the major AI crawlers it observed rendered JavaScript, while Googlebot does. Text that only appears after JavaScript runs may not reach them. See Can AI crawlers read JavaScript?
The four conditions
Tooru scores every page out of 100 on four conditions:
- Can AI find you? (19 points) robots.txt lets AI search crawlers in; sitemaps and canonical URLs show where pages are
- Can AI read the page? (29 points) The main text is in the server’s HTML; headings show the structure
- Does AI understand your business? (38 points) The page says what the company does and for whom; structured data identifies the organization
- Will AI use you in an answer? (14 points) Specific facts, sources, direct answers and visible dates
If a page fails conditions 1 and 2, the rest can’t help it. The weights and every check are in the methodology.
Where to start
- Read your robots.txt and decide, crawler by crawler, what you allow
- View the page source and confirm the main text is there
- Put a plain statement of what you do near the top of the home page, and add Organization structured data
- Answer common questions in two or three opening sentences, with sources and dates
Measuring it
AI answers change from one run to the next, so a single prompt is a weak measure. Fix what you can measure on the page first, then check real answers several times across several models.
Changelog
- 2026-10-09: Published