llms.txt, schema and structured data, explained in plain English

Why machines need labels

When you read a web page, you instantly know the phone number is a phone number and the list of towns is a service area. A machine reading the same page sees a lot of text and has to guess. Usually it guesses well. Sometimes it does not, and that is when an AI assistant tells a customer you are open on Sundays when you are not.

Structured data removes the guessing. It is a small block of code, invisible to visitors, that says in effect: this is a business, its name is this, its address is this, it is open these hours, it serves these towns, and here are its services. Search engines and AI tools read it directly.

Schema, in plain terms

Schema.org is a shared vocabulary for structured data, backed by the major search engines. When people say “schema markup,” they mean structured data written with that vocabulary, usually in a format called JSON-LD that sits in the page’s code.

You do not need to learn the whole vocabulary. A local business only needs a handful of types, and most of the value comes from getting those few right and keeping them accurate.

• LocalBusiness, or a more specific type like Plumber, Dentist, HVACBusiness or Restaurant: your core details
• Service: each service you offer, with a short description and area served
• FAQPage: questions and answers on a page
• Review and AggregateRating: where you display genuine reviews on your own site
• Organization: your logo, social profiles and official website
• BreadcrumbList: how your pages relate to each other

What llms.txt is

llms.txt is a proposal, introduced in 2024, for a plain text file at the root of your website (yoursite.com/llms.txt) written in a simple format called Markdown. The idea is to give AI tools a short, curated guide to your site: who you are, what you do, and links to the most important pages, without the menus, pop-ups and scripts that clutter a normal web page.

Think of it as a cover letter for AI. A Huntington accounting firm’s llms.txt might be a dozen lines: a one-paragraph description, the towns it serves, and links to its tax preparation, bookkeeping and small-business advisory pages.

The honest caveat: it is a proposal, not a standard, and the big AI companies have not publicly committed to using it. Some tools read it; how much it influences answers is unclear. It takes half an hour to write, so there is little downside, but do not expect it to move the needle on its own.

What matters most for a small business

If you have limited time, prioritize in this order. First, LocalBusiness schema with accurate name, address, phone, hours, service area and website. Second, FAQPage schema on the pages where you answer real customer questions. Third, Service schema for your main offerings. Then, if you like, an llms.txt file.

Accuracy beats completeness. Schema that says you are open until 6pm when your Google profile says 5pm creates exactly the kind of conflict that makes AI tools hesitant. Whatever you put in your structured data must match your website text, your Business Profile and your directory listings.

This is the same principle that drove one of our client projects. Mystic Journey Path, an e-commerce business, grew organic traffic 210% after we rebuilt its Shopify store as a custom Next.js storefront connected to Shopify by API and structured every product and page so search engines and AI assistants could read and surface it. A local service business is simpler, but the idea holds: label things clearly and consistently.

How to add it without a developer

Most website platforms make the basics manageable. WordPress has well-known SEO plugins that generate LocalBusiness and FAQ schema from forms you fill in. Squarespace and Wix add some business schema automatically and let you paste custom code into page headers. Shopify themes often include product schema, with apps to extend it.

For an llms.txt file, write it in any text editor and upload it to the root of your site. On some hosted platforms you may need a developer or support ticket to place a file at the root.

Whatever route you take, test it. Google’s Rich Results Test and the Schema.org validator will both tell you whether your markup is readable and flag errors.

• Write down your official business details once, and use that as the single source of truth
• Add LocalBusiness schema via your platform’s SEO tool or a code block
• Add FAQPage schema to pages with real Q&A content
• Validate with Google’s Rich Results Test
• Optionally write an llms.txt with a short description and links to key pages
• Re-check whenever your hours, address or services change

Common mistakes

The most common mistake is set-and-forget. A business moves from Hicksville to Plainview, updates its website footer, and leaves the old address in its schema for two years. Every time you change a detail, change it everywhere, including the hidden parts.

The second is marking up things that are not on the page, like FAQ schema for questions that do not appear on the visible page, or review stars from reviews that are not shown. Google treats that as misleading and can ignore or penalize it. Structured data should describe what is actually there.

The third is expecting miracles. Structured data makes your business easier to understand. It does not make a thin, vague website into a strong one. It works best layered on top of clear content and good reviews.