What Is llms.txt and Why Local Businesses Should Have One

In late 2024 a proposal went around for a new file called llms.txt. By 2026 it is an emerging standard that AI crawlers increasingly read first. Most local businesses still do not have one. This is what it is, why its absence costs you, and how to find out where you stand.

What llms.txt actually is

It is a file at the root of your domain (yoursite.com/llms.txt) that gives AI crawlers a curated summary of your business: who you are, what you do, where you do it, and which pages matter. Think of it as a briefing written for a language model instead of a human.

It is not robots.txt. Robots.txt tells crawlers what they may or may not access. llms.txt tells them what your site is about and how to read your business correctly. Different file, different purpose.

The point is simple. Without it, an AI engine has to crawl your site and infer who you are from your navigation and your prose. With it, you hand the engine the facts directly. One of those two situations produces a confident recommendation. The other produces a guess.

Why it matters for local search

AI search engines like ChatGPT, Perplexity, and Gemini build their answers by retrieving and synthesizing content from the web. Before any of them will put your name in a recommendation, they have to be confident about who you are, what you offer, and where you work. Confidence is the whole game, and confidence comes from having the facts in a structured, easy-to-read place.

For local businesses this matters more than for, say, a SaaS product. A local business lives or dies on NAP details, service categories, and service areas that an engine needs to be sure of before recommending you to a buyer who is about to call. When those facts are scattered across the site for the engine to assemble on its own, the engine assembles them wrong, or not at all, and recommends a business it understood more clearly. See why ChatGPT recommends your competitor and not you for the bigger pattern.

What its absence costs you

The recommendation set in AI search is tiny. A buyer asks “who is the best [trade] in [city],” and the answer names two or three businesses. There is no page two. You are in that set or you are invisible for that query.

When an engine cannot quickly and confidently read who you are, it does not flag you as uncertain to the buyer. It simply leaves you out and fills the slots with businesses it could read. The buyer never sees that you were skipped. They get three names and they dial one, and the choice was made before your site ever entered the picture.

That is the cost: not a worse ranking, but absence from the moment of decision. The job is booked with a competitor while you are still waiting for a call that an engine quietly routed elsewhere.

One picture of how it plays out

A buyer opens Perplexity and types “highly rated [trade] near [your city].” Perplexity pulls a handful of candidates, weighs how confident it is about each, and writes a short answer naming the two it understood best, with citations.

The two it names had given the engines a clean, structured account of themselves to read. Yours, where the same facts existed but were buried in page copy the engine had to crawl and interpret, did not make the cut. Not because you are worse. Because at the speed these answers get built, the legible business wins and the buried one is skipped.

You never see that query. You only see a slow month and wonder why.

Why DIY rarely closes it

The file itself is not the hard part. The hard part is everything around it. llms.txt is one signal in a stack: it does not replace structured data (JSON-LD schema), strong content, or citations. If the rest of your footprint is thin or inconsistent, the file alone changes nothing, and you will not know whether the problem is the file, the schema, the citations, or the way they conflict with each other.

Knowing which of those is actually keeping you out of the recommendation set, on your specific site and against the competitors winning your market, is a diagnosis. It is different for every business, and it is the part a template cannot give you.

Find out where you stand

You do not have to guess whether AI engines can read your business or are skipping past it. We run the real buyer questions for your market across ChatGPT, Perplexity, and Gemini, show you who is being handed the calls you should be getting, and put a number on what that gap is costing you. Then we install and run the fix so your business is the one the engines understand and recommend.

The diagnosis is free. The install is the work we do and keep optimized.

Book a call with our team. Related: the complete guide to AI search visibility, why ChatGPT recommends your competitor.

The short version

llms.txt is the file AI crawlers increasingly read first to understand a business. Most local businesses do not have one, and even a perfect one does nothing on its own if the rest of the footprint is thin. The cost of being unreadable is absence from a recommendation set that only holds two or three names. Finding out where you stand is free; closing the gap is the system we install and run.