Strategy & Positioning

Why llms.txt is not a strategy

llms.txt is a proposed file convention placed at /llms.txt that exposes a curated, markdown-formatted summary of site content for AI engines. Proponents pitch it as 'robots.txt for AI' or as a way to reach AI engines more directly. The skeptical consensus is now broad: Search Engine Journal, Similarweb, Webflow, and Generix Marketing's 2,500-site study have all published versions of the argument that llms.txt is not (yet) a meaningful AEO signal. This essay joins that consensus and adds a contribution - a 3-question test for any 'AEO trick' - so brands can systematically deprioritize emerging-standard busywork until fundamentals are solid. Major AI engines have not committed to reading llms.txt; brands that implement it while neglecting crawler access, schema, and content density see no measurable improvement in AI citation share. Implementing llms.txt is harmless busywork; treating it as an AEO strategy is a category error.

Updated 2026-05-06

Questions this guide answers

Does llms.txt help with AI visibility?
Should I implement llms.txt?
What is llms.txt?
Is llms.txt an AEO strategy?

Direct answer

llms.txt is a proposed file convention - placed at /llms.txt on a website - that exposes a curated, markdown-formatted summary of site content for AI engines to consume. Proponents pitch it as 'robots.txt for AI' or as a way to reach AI engines 'more directly.' In reality, the major AI engines (ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini) have not committed to reading llms.txt, adoption is concentrated among AI-curious developers rather than buyers' AI engines, and brands that implement llms.txt while neglecting basic crawler access, schema, and content density see no measurable improvement in AI citation share.

Implementing llms.txt is harmless busywork; treating it as an AEO strategy is a category error. If you have llms.txt on your roadmap, ship it after you've verified OAI-SearchBot, GPTBot, ChatGPT-User, PerplexityBot, and Googlebot access, fixed your schema, and shipped enough specific structured content to be cite-worthy.

Joining the skeptical consensus

The skeptical view of llms.txt is now broadly held across the SEO and AEO community. Recent published expressions of the same skepticism include Search Engine Journal (https://www.searchenginejournal.com/llms-txt-for-ai-seo/556576/), Similarweb's GEO team (https://www.similarweb.com/blog/marketing/geo/llms-txt/), Webflow (https://webflow.com/blog/llms-txt), and Generix Marketing's 2,500-site study (https://www.generixmarketing.com/learn/aeo/llms-txt-study/) which found no measurable correlation between llms.txt presence and AI citation lift.

This essay joins that consensus rather than claiming originality. SolCrys's specific contribution is the 3-question test below for any 'AEO trick' - a systematic way to filter emerging tactics so teams can deprioritize busywork until fundamentals are solid.

Where llms.txt came from

The proposal originated from developer-community efforts to make web content more easily consumable by LLMs. The core idea is reasonable: provide a clean, markdown-formatted summary of your site that AI engines can read without parsing JavaScript, navigating complex DOM trees, or stripping ads.

This is a pleasant idea. It is not a standard the major AI engines have actually adopted.

What the major AI engines actually use

In 2026, AI engines retrieve site content through one or more of: standard web crawlers (Googlebot, Bingbot, OAI-SearchBot, PerplexityBot, ClaudeBot), structured data in HTML (schema.org JSON-LD), sitemap.xml for crawl prioritization, search index integration (Bing index for ChatGPT, Google index for AI Overviews), and direct data partnerships for selected publishers.

None of these are llms.txt. None have publicly committed to reading llms.txt. None announce data-ingestion changes that include llms.txt. When OpenAI, Anthropic, Google, or Microsoft want to access content differently, they introduce a new bot (like OAI-SearchBot) and document it. They have not done that for llms.txt.

Why llms.txt feels strategic when it isn't

Three patterns drive teams to over-invest in llms.txt.

Pattern 1: Tactical clarity

llms.txt gives a clear, visible deliverable: 'we made a file called /llms.txt.' The fundamentals (crawler access audit, schema validation, structural refactoring) are messier. Teams reach for the visible deliverable to feel productive.

Pattern 2: First-mover bias

'If we're early on llms.txt, we'll have an advantage when it gets adopted.' This is the same logic that drove brands to invest in AMP pages, app indexing, and other deprecated standards. The cost-benefit of being early on an unadopted standard is bad.

Pattern 3: Vendor amplification

Some content marketing platforms have published 'guides to llms.txt' that imply it is a meaningful AEO move. These guides drive adoption, which drives traffic to the guide, which drives the platform's lead capture. The platform benefits from the meme; the brand implementing llms.txt does not.

When llms.txt does no harm

llms.txt does not hurt anything. Implementing it is cheap (a few hours of development), it does not break SEO, and if a major AI engine adopts the standard later, you'll be ready.

The harm is in the attention cost - when llms.txt occupies the AEO budget or the team's attention while the actual high-leverage actions (crawler access, schema, content density, third-party citations) are unaddressed. If you can implement llms.txt as a side task without delaying the fundamentals, do it. If implementing llms.txt is competing with fixing your robots.txt or shipping schema, fix the fundamentals first.

The 3-question test for any 'AEO trick' (SolCrys's contribution)

When someone pitches you a new AEO tactic - llms.txt, an AI-only schema, a special meta tag for ChatGPT - apply this test.

Has a major AI engine publicly committed to using this signal? (Not 'could' - has explicitly said they do.)
Are there documented cases where brands implementing this saw measurable citation lift?
Is the fundamental version of this problem - crawler access, schema, structured content - already solved at this brand?

Applying the test

If the answer to any of the three questions is no, deprioritize the new tactic. llms.txt fails questions 1 and 2 today. Most brands also fail question 3 - they have the new shiny tactic on the roadmap while their robots.txt blocks GPTBot.

What to focus on instead

The unsexy fundamentals that actually drive AEO citation.

Fundamental	Effort	Why it matters
Crawler access	About 1 hour	Confirm all major AI crawlers are allowed in robots.txt; the single biggest move many brands have not made.
Schema completeness	1 day per priority page	Run top 30 pages through Google's Rich Results Test; fix all errors; confirm Article, FAQPage, Product, Organization schemas.
Structural content density	1 week per priority pillar	H2 questions, FAQ blocks, direct-answer paragraphs, lists, and tables; the single most-cited optimization in our audits.
Bing indexability	1 day setup, ongoing	ChatGPT depends on Bing; many brands neglected Bing for years and pay for it now in ChatGPT visibility.
Third-party citations and community presence	90-day projects each	Editorial outreach, Reddit/forum engagement, G2/Capterra reviews; slow-compounding but durable.

The technical-honesty principle

This essay is part of SolCrys's technical-honesty stance: we will not promote AEO 'tricks' that have no evidence base, even when the trend is to do so. The category will be better when more vendors do this. The buyer is best served by clear distinctions between 'this is documented to work' and 'this is interesting and might matter someday.'

How to use this essay

If your team or your agency is recommending llms.txt as an AEO strategy: ask which AI engine has documented using it, ask for a measurable case where it produced citation lift, verify the fundamentals are addressed first, and then ship llms.txt as an addendum if all three are satisfied.

If you are deciding what to invest in for AEO, start with the fundamentals - crawler access, schema, content density, Bing indexability, third-party citations - before chasing emerging standards.

FAQ

Will major AI engines start using llms.txt?

It is possible. It is also possible they will introduce different signals. Standards adoption depends on what the engine providers prioritize for their own retrieval pipelines. Bet on documented signals, not on hopeful proposals.

Does it hurt to have llms.txt?

No. It is technically harmless. The risk is opportunity cost - time spent on llms.txt while crawler access or schema is broken.

Aren't there variations like /llmstxt.txt or /.well-known/llms?

Several variations have been proposed. None are adopted by major AI engines. The same logic applies.

What is the difference between llms.txt and robots.txt?

robots.txt is a long-established standard that controls crawler access. All major AI crawlers respect it. llms.txt is a new proposal that has not been adopted by major AI engines. They are not equivalent, even though the names suggest parallelism.

What about meta tags or HTML attributes specifically for AI?

Same skepticism applies. The schema.org standard is widely adopted by AI engines. Other proposed AI-specific meta tags are mostly speculative. Stick with schema.org until major engines document otherwise.

Should I take this essay's advice if SolCrys benefits from it?

The honest answer is yes - if SolCrys is wrong about llms.txt, the AEO platforms that promoted it benefit and SolCrys looks behind. The advice in this essay reflects what we believe to be the empirical reality, not a competitive jab. Test our claim against your own data and make your own call.

Related guides

Technical Readiness

Find out where your brand is missing, miscited, or misrepresented.

SolCrys maps high-intent prompts to mentions, citations, answer accuracy, and content gaps so your team can prioritize the next pages to ship.

Get a free audit