Minimal Technical AEO Checklist
Make pages readable, connected, structured, and ready for agent actions
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 98%
This checklist treats technical AEO as an extension of proven crawlability fundamentals. Important pages need strong internal links, content that renders when fetched, indexable access, and relevant schema or markup. Teams should also eliminate ordinary failure modes such as redirect loops, inaccessible content, or progressively loaded sections absent from the initial response. As answer engines gain the ability to perform actions, forms and buttons may require additional machine-readable markup. The method rejects unsupported rituals such as adding LLMs.txt or converting ordinary pages to Markdown merely because those ideas sound plausible. Its output is a short technical backlog tied to observable machine access rather than a long list of speculative optimizations that distracts from content quality and authority.
Origin
Extracted from Marketing Against The Grain
Core principles
- 01Crawlability matters more than novel AEO files
- 02Technical fundamentals serve both SEO and AEO
- 03Special formats are unnecessary without evidence
- 04Action markup becomes relevant when agents transact on sites
How to run it
- 1
Audit internal discovery
Confirm that important pages receive clear internal links from relevant, crawlable locations.
Pro tip Inspect whether high-value pages are orphaned or buried deeply in the site structure.
Watch out A sitemap alone does not provide the contextual relationships of good internal linking.
- 2
Verify initial rendering
Fetch each important page as a crawler would and confirm that its substantive content appears without delayed client-side interaction.
Pro tip Compare server output with the fully rendered browser page.
Watch out Progressively loaded sections may be absent when a search or answer engine fetches the page.
- 3
Confirm indexable access
Check robots directives, status codes, canonicalization, redirects, and other controls that affect machine access.
Pro tip Test representative templates rather than checking only the homepage.
Watch out Redirect loops and accidental exclusions can nullify otherwise strong content.
- 4
Apply relevant structure
Use schema and semantic markup that accurately describes the page and its entities.
Pro tip Prefer standards already consumed by major crawlers over speculative formats.
Watch out Markup must match visible, truthful page content.
- 5
Prepare for agent actions
Monitor emerging standards for marking up forms, buttons, and transactions that autonomous agents may need to operate.
Pro tip Prioritize high-value actions such as booking, purchasing, or requesting information.
Watch out Do not implement immature conventions as if support were already universal.
- 6
Reject unsupported additions
Require evidence before adding LLM-specific files, Markdown transformations, or other special technical layers.
Pro tip Test whether a proposed change alters discovery, citation, or conversion before standardizing it.
Watch out Being crawled does not prove that a file influences an answer.
In the wild
A product page places crucial integration details in a section loaded only after the visitor scrolls. The team changes the template so the details appear in the initial rendered response, links the page from the integration directory, and applies accurate structured markup.
→ Both traditional crawlers and answer engines can consistently retrieve the evidence needed to describe the integration.
Common mistakes
Treating LLMs.txt as established
A crawler requesting a file does not demonstrate that the file affects retrieval, ranking, or answers.
Converting pages to Markdown unnecessarily
There is no general evidence that major LLM companies require ordinary web pages to be transformed into Markdown.
Overlooking basic crawl barriers
Speculative AEO work cannot compensate for missing links, failed rendering, redirect loops, or blocked pages.
Is it for you?
Best for
Website teams auditing whether search engines and LLMs can reliably discover, read, interpret, and eventually act on their pages.
Not ideal for
Sites with severe platform-specific rendering defects that require a deeper engineering investigation beyond a general checklist.
From the transcript
“Internal links, schema, markup, and make sure the LLM can read your page. That's really all that there is.”
“But LMS.txt, as John Mueller from Google said, as far as I know, no one uses it, we don't use it.”
From the episode
‘My Data Proves SEO is NOT Dead’ + How to Rank #1 on Google & AI