Lesson 1 · 16 min
AEO (optimizing for answer engines) and GEO (optimizing for generative AI) have become a package sold with a file, markup, and a promise. In May 2026, Google published a guide that takes much of that apart. This lesson uses that guide as a yardstick to decide where your time is worth spending.
Start the lesson
Goal: separate what Google, Bing, and the AIs officially say from what's a fad, and choose where to spend your time.
Click each block.
Spend your time on the first three blocks. The fourth has no source to support it as a factor.
Click each one. All were read on the official pages on 10/19/2026.
You don't. Google's AI features page says you don't need to create machine-readable files, text files for AI, or special markup, and that there is no special schema.org to add.
Structured data is still useful when it matches the visible text, just as it was before.
The largest public study (Ahrefs, 06/15/2026, 137 thousand domains) found 97% of valid llms.txt files had no requests at all in the month. Google's guide says you can ignore it.
Honest read: llms.txt is optional and cheap, useful as an index for coding agents, and not a citation lever.
No source supports that length as a factor. A direct answer at the start of a section is good for readers; treat it as clarity, not an AI trick.
It matters twice: Bing feeds Copilot and has the only official AI citations dashboard. And a Seer study found 87% of ChatGPT citations in Bing's top 20; it's a small study, not a statement from OpenAI. Lesson 5.
Copy it and fill it in with the promises you've already heard. One row per promise.
# AEO/GEO promises triage (date: dd/mm/yyyy) | Promise I heard | Official source that checks it | Verdict | |---|---|---| | "llms.txt gets you into ChatGPT" | Google: AI guide (can ignore it); Ahrefs 06/15/2026: 97% no reads | optional, cheap, not a lever | | "special schema for AI" | Google: AI features (no special schema exists) | fad | | "content with your own data" | Google: AI guide (non-commodity) | foundation | | <promise 4> | <official page> | <foundation / optional / fad> | | <promise 5> | <official page> | <foundation / optional / fad> |
List 3 things my business knows that no competitor or AI could write on their own: 1. a number only I have (e.g.: real average delivery time, price charged, measured result) 2. an experience only I have lived (e.g.: a case, a mistake, a before and after) 3. an opinion I defend with a reason For each one: which page on my site already says it? If none does, which page would be the owner page?
In July 2026 INEMA measured its own citation: it appeared in 0 of 48 neutral answers on Claude and 0 of 4 on Gemini (52 neutral measurements). After that it published 54 knowledge sheets and started measuring in Search Console and Bing.
The lesson: the problem wasn't a missing file. It was missing content that answered the neutral questions on the topic.
Links at the end of the lesson.
A made-up conversation for practice. Pick the answer you would give; the comment appears right after.
Choose your answer. The comment appears right after, and the conversation continues based on your choice.
Use the template from the block above.
Fill in the triage table with 5 AEO/GEO promises you have already heard (from a client, an ad, a video). For each one, open the official page and write the verdict.
Done when: the triagem-geo.md file has 5 rows, each with the official page and a verdict (foundation, optional, or fad).
Answer the question from the second block for your site or a client's. For each item, point to the page that already covers it or the one that should.
Done when: you have 3 items (number, experience, opinion), each with the URL of the owner page or the note "page missing".
The explanation appears after your answer.
Question 1 of 4
According to Google (May 2026), what do you need to do that's special to appear in AI Overviews and AI Mode?skill
Question 2 of 4
Which of these is non-commodity content?
Question 3 of 4
What can you say, with a source, about llms.txt?
Question 4 of 4
A paid tool shows your "official AI visibility score." How do you read it?
Choose an answer. The explanation appears after you try.