Deep dives
What is llms.txt, and do you need one?
An agency workflow for deciding whether llms.txt is useful, assigning its ownership, checking it after deployment, and reporting the limits clearly.
A client asks whether they need an llms.txt file because another agency recommended one.
The useful response is not a blanket yes.
It is a documented decision about whether the client has a maintained set of pages an agent should be able to discover and whether someone will own the file after launch.
llms.txt is a markdown file served at a site's root that lists important pages with short descriptions. It is a convention, not a confirmed ranking mechanism, and it does not guarantee that an AI engine will fetch, cite, or prioritize a site.
Decide whether the client has a useful file to publish
An llms.txt file is worth considering when a client has a maintained product, documentation, or knowledge surface with a clear set of canonical pages. It is less useful when the site is small, the important pages are unsettled, or no one will verify the file after route changes.
An agency practitioner can use this decision table:
| Question | What to check |
|---|---|
| Is there a stable set of canonical pages? | Product, pricing, documentation, API, or help pages that answer real reader questions. |
| Does each page have a clear owner? | A team can correct stale links and descriptions. |
| Can the file be served at a predictable URL? | The client can maintain a root-level file in its deployment workflow. |
| Is crawler access separately reviewed? | The file does not make an otherwise inaccessible site readable. |
| Can the client accept the limit? | The file is not a promise of citation, traffic, or ranking. |
Turn the decision into a client handoff
The agency's work is to make the recommendation traceable. Record the decision, the selected pages, the owner, and the next validation date in the client plan.
A technical implementer can generate or serve the file. The agency practitioner should check that the content reflects the client's current positioning and links to the pages the client wants readers and agents to verify.
Keep the file concise. It should identify the site and describe a bounded set of maintained pages, not reproduce a sitemap or every URL in the CMS.
Verify the file after deployment
The practical verification is simple. Check that the root URL returns the intended text content, then verify every linked URL resolves to the right current page.
Also check whether the site's relevant pages can be fetched according to the client's crawler-access policy. A valid llms.txt file does not overcome a blocked crawler, a JavaScript-only empty response, or stale linked pages.
Maverank's readiness audit treats this as a checkable condition and reports link or access findings as a fix list. The result still needs a human decision about whether the file's content represents the client accurately.
Report the limits without weakening the recommendation
A client report should say what was changed and what was not established. For example: the file is present, its links resolve at the time of the check, the listed pages match the agreed canonical set, and no claim is made that it will change AI-engine citations.
This is more useful than claiming that the file is a visibility tactic. It gives the client an auditable change and a date for reassessment as conventions and provider documentation evolve.
Frequently asked questions
Does llms.txt make a site appear in AI answers?
No evidence establishes that outcome. Treat the file as a maintained discovery convention, not a guarantee of citation, ranking, or traffic.
Who should own a client's llms.txt file?
The client should own its content and canonical-page decisions. An agency can recommend, verify, and report the work, while a technical owner maintains deployment and route changes.
Does llms.txt replace robots.txt or a sitemap?
No. Robots rules, sitemaps, server responses, and rendering each have separate roles in how a site is discovered and accessed.
Check the file in the wider readiness workflow
Run a free check to review llms.txt alongside crawler access and other readiness conditions. For the service workflow around those findings, see how agencies deliver AI visibility work.