Retrieval and usability, control 17 of 18
R-17 Do not impose unsupported AI markup or chunking rituals
Do not make llms.txt, proprietary AI schema or arbitrary micro-chunk lengths mandatory for Google Search. Another system may use such files for a separately documented purpose.
- Priority
- High priority.
- Hard gate
- No
- Change class
- AUTO_WITH_TEST: A candidate requiring relevant tests.
- Quality bucket
- Retrieval and usability
Why it matters
Content chopped into artificial fragments or written to match imagined prompts is harder for people to read and gains nothing on its own. Google's guidance on generative AI features points back to ordinary crawlability, clarity and helpful content rather than special formats. Chasing unproven tricks also pulls time away from the basics that actually make a page eligible.
Evidence to review
- Platform requirement or clearly labelled hypothesis, implementation purpose and measured outcome.
How to implement
- Ask which platform consumes the proposed format and what evidence supports the benefit.
- Prioritise accessible, well-structured useful content.
- Keep optional agent documentation separate from ranking claims.
- Test experiments without degrading human comprehension.
How UltimateSEO checks this
Partly automated. The scan records information signals only, so this control never becomes PASS or FAIL on its own.
The check looks at: llms.txt presence is reported as information only and never as a failure; the team confirms that no AI-specific hack is being treated as a base requirement.
States used: PASS, FAIL, UNKNOWN, NOT_APPLICABLE or BLOCKED. UNKNOWN is never shown as a pass. How states work.
Myth guards that reference this control
Sources
Part of The UltimateSEO Framework v4.3, catalogue dated 2026-10-14. Evidence states, control states and the publish gate are explained on the hub.