Red Mage
Book a call

ResourcesUpdated

What Generative Engine Optimization (GEO) Actually Is

Most writing about GEO is selling something. Here is the version without the pitch.

Key takeaways

  1. 1GEO makes your site easy for AI answer engines to quote as a source.
  2. 2What reliably helps: pages that rank in normal search, quotable sentences, and facts that match everywhere.
  3. 3Cheap to add, benefit unconfirmed: /llms.txt and JSON-LD aimed at AI engines.
  4. 4What matters less than claimed: keyword density and backlink counts.
  5. 5There is no Search Console for ChatGPT. Track brand searches, chat referrals and monthly spot checks.

GEO in one paragraph

Generative Engine Optimization is making your site easy to cite for AI answer engines: ChatGPT search, Perplexity, Claude with web access, Google's AI overviews, Copilot. It overlaps with classic SEO. The goal is to be the page an answer quotes.

What changed from SEO

GEO compared with classic SEO
Matters more nowMatters less than claimed
Sentences a model can quote as they areKeyword density: models read meaning, not term counts
Being indexed and ranking in regular search, which most answer engines draw fromBacklink volume: a few trusted citations beat many links
The same names, dates and numbers on every pageSpeed micro-tuning: still matters for people, not for citation

What is unproven

/llms.txt is a proposed file, not a standard. Google's John Mueller has said no AI system currently uses it (opens in a new tab), and no major answer engine has confirmed reading it. JSON-LD helps classic search results; there is no public evidence that it changes AI citations. I still add both because they take an hour, but don't pay anyone extra for them.

Google's AI Overviews and AI Mode are part of Search. Per Google's documentation (opens in a new tab), a page only needs to be indexed and eligible for a snippet to appear there. The Google-Extended robots.txt token does not change that; it controls use in Gemini and Vertex AI.

The checklist

I add these to every site I ship.

  • Pages indexed in Google and Bing. Check Search Console and Bing Webmaster Tools first; Copilot answers from Bing's index.
  • FAQ answers written as a direct answer to the question
  • robots.txt that allows search crawlers (Googlebot, Bingbot, OAI-SearchBot, PerplexityBot, Claude-SearchBot). Allowing training crawlers (GPTBot, ClaudeBot, Google-Extended, Applebot-Extended) is a separate choice.
  • JSON-LD on every page: Organization, Article, Service, FAQPage, Person
  • /llms.txt and /llms-full.txt, as a cheap bet rather than a known signal
  • The same names, dates and numbers on every page, so models can cross-check them
  • Citations on two or three trusted outside pages: a platform profile, a conference page, a press mention

What to track

  • Searches for your brand name, as a sign answer engines are mentioning you
  • Referrals from chatgpt.com, perplexity.ai and claude.ai. Small numbers, still useful.
  • A monthly spot check: ask Perplexity "who builds websites for nonprofits?" and see if you show up

Common questions

Should I block AI crawlers instead?
Only if you publish content that crawlers monetize without paying you. For a service business, being cited is the goal.
Will GEO replace SEO?
No. They overlap, and classic search still drives most discovery. GEO is the same content, packaged so answer engines can quote it.
Will this still matter in two years?
The tools will change. Clean, structured content that machines can read has mattered since the first search engine and will keep mattering.

By Andrés Sepúlveda Morales. Top Independent on Contra, rated 4.99 (opens in a new tab) by clients and collaborators, and a speaker on AI in English and Spanish.

Check your site. Audit from $500.