Think of your llms.txt file as a formal biography you're submitting directly to AI — not a sales pitch, not a brochure. A biography. And like any document that gets scrutinized, one wrong detail can stop the whole thing cold.
Until recently, llms.txt wasn't part of the conversation for most local businesses. That's changing fast. Within the AI Ranking Factors framework, it's emerging as a structured signal that helps AI assistants evaluate who a business is, how stable it is, and whether it can actually deliver for the person asking the question.
Here's how to build it so it works.
When someone asks an AI assistant to recommend a plumber, a restaurant, or an accountant, the AI isn't just pulling a name from a list. It's running a quick trust evaluation. Can this business be found? Has it been around? Does it serve the public reliably? Does the information about it hold together across multiple sources?
Your llms.txt file speaks directly to that evaluation. It's not for humans browsing your site. It's a clean, factual document that gives AI a single place to confirm your identity and credibility — without wading through marketing copy about your "passion for excellence."
Strip the fluff. Lead with facts that can be verified.
Large chains and big-box retailers have an inherent advantage in AI evaluations: they appear consistently across dozens of directories, have stable domains, show up in news coverage, and maintain uniform information at scale. Our analysis suggests they pick up significantly more weight on stability signals than the average small business.
You can't out-distribute a national chain. But you can make sure your information is airtight. That's where llms.txt earns its place.
This is the foundation. Your business name must match exactly — not approximately — across every place it appears:
One character difference — an ampersand instead of "and," a missing comma, an old street address — can break the chain of verification AI uses to confirm you're legitimate. Check every instance before you publish this file.
Your address in llms.txt needs to match your address everywhere else: your website, your Google Business Profile, LinkedIn, any trade association directories. If you've moved and forgot to update an old listing, find it and fix it before you build this document. A conflicting address signals instability, not just sloppiness.
AI will often cross-reference your domain registration date to estimate how long you've been operating. If your business predates your current domain, say so — explicitly, with a date. A brief founding story is appropriate here, but keep it factual: who started the business, when, and why it exists.
If there's a verifiable external record — a news article, a local business journal mention, a chamber of commerce announcement — link to it. External corroboration carries weight in our index.
Name the current owners. If they have professional profiles on LinkedIn or elsewhere, link to them. If they've been mentioned in press coverage, include those links. If ownership has changed since founding, note the transition clearly.
For licensed trades — contractors, healthcare providers, financial advisors — include your license number and a direct link to the public verification source where AI can confirm it. This is one of the strongest trust signals you can provide.
You don't need to list every employee. Focus on the ones whose credentials matter to what you're selling:
For each relevant staff member, include any professional association memberships, certifications, or notable recognition — with links where those credentials can be independently confirmed.
Is your business a member of the local chamber of commerce? A trade association? A professional licensing board? List them, and link to the membership directory where your business appears. These associations function as endorsements in AI's evaluation — they signal that you're embedded in a professional community, not operating in isolation.
Write it like a proposal to a careful reader who has no reason to trust you yet. Plain language. No hyperbole. Every claim you make should point to something that can be checked.
Think of it as the document you'd hand to a due-diligence reviewer before a business deal — not the brochure you'd hand to a first-time customer.
Building an llms.txt file isn't a one-time project. As ownership changes, staff turns over, licenses renew, and your business evolves, this document needs to reflect reality. AI doesn't reward what you used to be. It evaluates what you demonstrably are right now.
Get the facts right. Keep them current. Make them verifiable. That's the entire job.
— J. Brent Tuttle