Generative Engine Optimization (GEO): Get ChatGPT to Recommend Your Business
Summary
A YouTube class walks non-technical business owners through building a second, machine-readable version of their existing website that lives at the same address as the human version. A small router on Cloudflare inspects the requesting agent: normal visitors and Google get the standard site, while crawlers such as GPTBot, ClaudeBot, and PerplexityBot get pages written to be cleanly quotable. The build is planned and shipped with Claude Code rather than hand-written, and the session ends with live checks plus a weekly loop for adding answers to questions people actually ask.
The canonical URL points at a public GitHub repository, claude-seo, which packages the underlying capability as a reusable SEO skill for Claude Code covering technical SEO, schema, GEO and AEO, backlinks, local SEO, and reporting. The repository, its changelog, and its releases were fetched and confirm the toolkit exists and is maintained in the open.
Demonstration by Eric Grows Companies, embedded from YouTube. AIOTruth did not reproduce the test.
Why it matters
The argument here separates two things that usually get collapsed: being discovered and being used. A model has to find your page through search, then successfully lift a passage from it, and a page that renders fine for a person can still be difficult for a crawler to extract a clean answer from. That reframes the work from ranking to liftability, and it makes the Cloudflare crawler dashboard a direct feedback signal: you can see which AI agents requested your pages rather than guessing. The approach also raises a live question anyone copying it has to answer, since serving different content by user agent is the shape of cloaking, and the class addresses it by asserting both versions show the same substance.
Source
- github.com/AgriciDaniel/claude-seo the item itself, published by Eric Grows Companies, created by Eric Grows Companies
- github.com/AgriciDaniel/claude-seo/blob/main/CHANGELOG.md Public source code or repository, first party
- github.com/AgriciDaniel/claude-seo/releases Public source code or repository, first party
- github.com/features/copilot Public source code or repository, first party
- github.com/features/ai/github-app Public source code or repository, first party
- docs.github.com/en/site-policy/github-terms/github-terms-of-service Public source code or repository, first party
- github.com/SomethingWithComputers/pixoo Public source code or repository, first party
- github.com/cyanheads/pixoo-toolkit Public source code or repository, first party
- github.com/blackbeardlabs/ds4x_adaptive_dspark_production_bundle Public source code or repository, first party
- youtube.com/watch?v=nxRY-s7JmxIwhere AIOTruth found it
What was checked
- read the item page (200)
- https://github.com/AgriciDaniel/claude-seo/tree/main/docs: does not mention the item, tier 6 (Public source code or repository)
- https://github.com/AgriciDaniel/claude-seo/blob/main/CHANGELOG.md: supports the item, tier 6 (Public source code or repository)
- https://github.com/AgriciDaniel/claude-seo/releases: supports the item, tier 6 (Public source code or repository)
- https://github.com/AgriciDaniel/claude-seo: supports the item, tier 6 (Public source code or repository)
- https://github.com/features/copilot: supports the item, tier 6 (Public source code or repository)
- https://github.com/features/ai/github-app: supports the item, tier 6 (Public source code or repository)
Limitations
- Everything checked traces back to a single domain, so this is one party's account.
- The demonstration was not independently reproduced by AIOTruth.
AIOTruth judgment
Relevance scored 10 because the mechanics of whether an AI system can extract and cite a page sit at the center of what this site covers. Usefulness at 8 reflects a concrete, followable build with a public repository behind it rather than advice. Evidence at 7 and the Supported status come from what could actually be checked: the repository, its changelog, and its releases are real and first party, and the presenter states plainly that AI visibility is a slow build measured in weeks rather than a switch. Originality at 5 is the honest read, since the dual-serving pattern and the found versus used framing are both circulating widely. The limiting factor is that everything traced back to a single domain, so this is one party's account, and the demonstration was not independently reproduced.
How this score was calculated
| Dimension | Weight | Score | What produced it |
|---|---|---|---|
| AIO relevance | 30% | 10 | generative engine optimization, geo, ai visibility, aeo, ai crawler; 4 scope question(s) matched |
| Usefulness | 30% | 8 | 3 artifact(s), 6 actionable marker(s), 0 measured figure(s) |
| Evidence | 25% | 7 | Public source code or repository; 5 verified source(s) across 1 domain(s) |
| Originality | 15% | 5 | no original testing found; closest archive match 0 |
aioRelevance x 0.30 + usefulness x 0.30 + evidence x 0.25 + originality x 0.15. The rubric is published in full on the Editorial Method page. Scoring is deterministic: the same item scores the same on every run.
Watch the demonstration Open the source
Found something wrong here? AIOTruth corrects material errors openly. Have something we should review? Submit a find.