Skip to content
verastack/labs
Launch kit Start a project
← all notes

18 October 2026 · 3 min read

GEO is the new SEO: what changes and what doesn't

Generative engine optimisation, AEO and SEO, untangled. What actually changes when buyers ask ChatGPT instead of searching, and the SEO work that still carries it.

For twenty years, getting found meant ranking: be one of ten blue links and hope for the click. Now a growing share of buyers never see the links. They ask ChatGPT, Perplexity or Google's AI Mode which tool to use, read one written answer, and shortlist whoever is named in it.

Getting named in that answer has picked up two names. AEO, answer engine optimisation, came first. GEO, generative engine optimisation, is the one founders search for now, after the engines that generate their answers. They describe nearly the same work. Here is what that work is, and how much of it is still plain SEO.

What doesn't change

Most of GEO is SEO done properly, for a reader that happens to be a model.

  • Crawlers have to get in. If robots.txt or your CDN's bot protection turns away OAI-SearchBot or PerplexityBot, nothing else matters. Our check caps the score at 40 when that happens, however good the rest of the page is.
  • The words have to be in the HTML. Many AI crawlers do not run JavaScript. A homepage that only says what you do after an animation loads is blank to them. In our YC Summer 2026 study, 30 of 226 homepages sent 50 words or fewer before JavaScript ran.
  • Titles, descriptions, headings and a sitemap still tell an engine what a page is and where the rest of the site lives.

If your SEO is broken, your GEO is broken. That part is not new, it just got punished harder.

What changes

The unit is a passage, not a page. A search engine ranks pages. A generative engine lifts a sentence or two and builds an answer around it. Pages that state the thing plainly, in one quotable sentence, get used. Pages that circle it in brand language get skipped.

Being cited matters more than being ranked. When a model names three tools, it leans on what other sites say: review sites, directories, comparison posts, Reddit threads, the press. Your own site is one voice in that answer, not the whole of it.

Facts beat adjectives. Prices, numbers, who it is for, how it compares. A model can quote "from $49 a month, for teams of 5 to 50"; it cannot quote "powerful and intuitive".

Structured data stops being optional. Organization data tells an engine which company you are among the namesakes. Product or SoftwareApplication data tells it what you sell. A marked-up FAQ hands it question-and-answer pairs to quote. In our studies, these were the gaps on most homepages, from Indian D2C stores to established SaaS.

There is no rank to check. You cannot look up position three for a prompt. The only way to know whether you are named is to ask the engines the questions your buyers ask, write down who gets named, and ask again after every change.

AEO vs GEO vs SEO, in one table

SEO AEO and GEO
Goal A high place in a list of links Your name inside the written answer
Reader A search crawler A crawler, then a model writing an answer
Unit The page The passage and the fact
Proof Rankings and clicks Being named, checked by asking the engines
Foundations Crawlable, fast, well-labelled pages The same, plus structured data and citations

Where to start

Start with what a model can read today. Our AEO and GEO check runs sixteen checks on one page in about ten seconds and gives you the three fixes that matter most. It's free.

The page is step one. Whether ChatGPT names you or a competitor comes down to the questions your buyers ask, the sites the answers draw on, and pages worth quoting. That is the audit we run, and the work we do. Tell us about your site.

How does your site read to AI search?

Across 226 YC Summer 2026 homepages, the median AEO and GEO score was 66. Where does yours land?