Free AI SEO Course: A Syllabus Built From Primary Docs

You don't need a paid platform to learn AI search. The companies that run the engines publish how their crawlers and AI surfaces work. A small set of studies tests the tactics that agencies sell. Reading both in order gives you two lists: facts that decide whether an engine can reach your page, and tactics with no evidence behind them.

One finding frames the whole syllabus. A July 2026 critical survey of 45 GEO studies (arXiv:2607.14035) concluded that already-retrieved content can change its citation, but no reviewed technique showed a stable, cross-platform causal effect on discoverability.

It also helps to keep four stages apart:

• Parametric: what the model knows with no web access. Nobody controls it this quarter.
• Access: whether a crawler may fetch and store your page. You and your CDN control this.
• Retrieval: whether your page enters the answer's context for a query. The engine controls this.
• Generation: which retrieved pages get used and cited. The model controls this.

If you want a broader method for working from primary sources, see how to learn AI SEO from primary sources.

Read these in sequence. Each module builds on the last.

1. Google, AI Features and Your Website (developers.google.com/search/docs/appearance/ai-features). The key line: a page must be indexed and eligible to show with a snippet to appear as a supporting link in AI Overviews or AI Mode, with no additional technical requirements.
2. Google's Guide to Optimizing for Generative AI Features (developers.google.com/search/docs/fundamentals/ai-optimization-guide). Google says you don't need AI text files, special markup or AI-specific writing, and that there's no ideal page length.
3. Google's robots meta tag spec and common crawlers page. Learn noindex, nosnippet, max-snippet and data-nosnippet. Note that Google-Extended governs Gemini Apps training and grounding, not Search inclusion.
4. OpenAI's crawler overview (developers.openai.com/api/docs/bots). GPTBot governs training. OAI-SearchBot governs ChatGPT search. OpenAI states each setting is independent of the others.
5. OpenAI's ChatGPT search help page. Eligibility needs OAI-SearchBot allowed and a host or CDN that lets OpenAI's published IP ranges through (openai.com/searchbot.json).
6. Anthropic's crawler article. ClaudeBot (training) is separate from Claude-SearchBot (search quality) and Claude-User (user-triggered).
7. Perplexity's crawler page. PerplexityBot indexes for search. Perplexity says Perplexity-User generally ignores robots.txt.
8. Meta's web crawlers page. It carries the only first-party statement in this set tying access to citation: allowing Meta-WebIndexer helps Meta cite and link to your content in Meta AI.
9. The studies. Read Vercel's AI crawler study, the original GEO paper, C-SEO Bench, the 252,000-trial factorial, and Ahrefs' schema test, in that order.

For the Google half in more depth, read Google's helpful content guidance and E-E-A-T.

Reading is half the course. The other half is checking your own site against what you read. These commands work from any terminal.

curl -s https://example.com/robots.txt
curl -s -o /dev/null -w "%{http_code}\n" -A "OAI-SearchBot" https://example.com/pricing
curl -s https://example.com/pricing | grep -i "per month"

What each one does:

1. Read robots.txt per bot. Check rules for OAI-SearchBot, PerplexityBot, Claude-SearchBot and Meta-WebIndexer separately from training bots like GPTBot and ClaudeBot. A blanket AI block often removes search eligibility by accident.
2. Request a page as a crawler. A 200 is a pass. A 403, 429 or a challenge page means your firewall or CDN is turning the bot away. Use the full user-agent string from each operator's page. Your own IP is not the crawler's IP, so a firewall that verifies IPs may treat the real bot differently.
3. Check the raw HTML. Vercel found ChatGPT and Claude crawlers fetch JavaScript files but don't execute them. If your price or product name only appears after JavaScript runs, those crawlers can't read it.
4. Check for suppression. Search your HTML for data-nosnippet and robots meta tags with nosnippet, max-snippet:0 or noindex.
5. Check Search Console. Google's generative AI inclusion control is opt-out by default and child properties can inherit it. See whether Search Console is hiding you from AI.

The full routine lives in the AI crawler access audit playbook.

Every claim you meet in a course, blog or pitch fits one of two bins. Use the primary sources to decide which.

One caution on overcorrecting. "Schema is useless, remove it" isn't supported either. Google says existing markup does no harm. It just isn't a lever for AI answers.

The positive signal is narrow: stating a specific fact, like a price or a one-line definition, on a fetchable page gives an engine a sentence it can quote. Topical relevance to the query also holds up across C-SEO Bench, the 252,000-trial factorial and the 45-study survey.

Primary docs have limits, and a good course names them.

• Google's statements are assertions, not measurements. They are still the best evidence for Google surfaces.
• Retrieval is per query and discretionary. AI Overviews often don't trigger, and ChatGPT decides when to search. Any visibility number mixes answers given with and without live retrieval.
• Rank doesn't exist to track. As of October 2026, Google's AI report shows impressions only, with no position. See why there's no AI rank to track and how to see AI Overview impressions in Search Console.
• Studies age fast. Ahrefs found 37.9% of AI Overview citations also appeared in the first 10 results in March 2026, but it changed methods between studies, so treat older comparisons as different instruments.
• Adversarial content exists. Research on search agents shows injected instructions can sway answers once that content is retrieved. It's a risk to watch on third-party pages, never a tactic to use.

The next move: treat access as the part you own, and the answer itself as the part you don't.

Once you've read the syllabus, you can automate the repetitive checks. The AI Crawler Checker requests a page as each AI crawler and shows which ones your firewall turns away, which covers step 2 above without a terminal.

To see the other end, the free checker on this site runs up to three buyer questions and shows whether AI answers name you. It's a snapshot, not continuous tracking. Use it after the access checks, so a missing mention points you at something you can actually fix.

Do I get a certificate from this free AI SEO course? No. This is a reading order, not an accredited program. The value is that every module is the operator's own documentation or a published study, so you can verify each claim yourself. If a hiring manager or client asks, you can point to the exact Google, OpenAI or Anthropic page behind your advice. That tends to carry more weight than a badge from a course that summarises those pages secondhand.

Should I remove schema markup from my site? No. Google says existing structured data does no harm, so removing it isn't supported by the evidence either. The honest position is that schema isn't a lever for AI answers. Ahrefs' matched test found no major uplift on any platform, and a 4.6% fall in AI Overview citations. Note its limit: every tested page was already heavily cited, so whether schema helps a page enter the answer is untested.

Does allowing GPTBot help ChatGPT find my site? No. OpenAI's crawler page describes GPTBot as a training crawler only and states that each bot setting is independent. ChatGPT search uses OAI-SearchBot. To be eligible for ChatGPT search, allow OAI-SearchBot in robots.txt and make sure your host or CDN accepts traffic from OpenAI's published search bot IP ranges. Blocking GPTBot does not remove you from ChatGPT search.

Which free course module should a busy founder read first? Read Google's AI features page and OpenAI's crawler overview first. Together they explain index eligibility, snippet controls and the split between training and search bots. Then run the robots.txt and crawler request checks on your own site. Those two readings and two commands cover the access gate, which is where the strongest evidence sits. The studies can wait until you need to evaluate an agency pitch.