Does Schema Markup or llms.txt Help You Show Up in AI Search?

Schema markup helps a little and llms.txt, as far as anyone can confirm, does not help at all for local AI visibility. Schema is good hygiene that helps Google understand your business, so it is worth doing properly, but it is not a switch that makes ChatGPT or Gemini recommend you. The bigger levers are consistent listings, genuine reviews and being mentioned on sources the assistants actually read.

That is not the answer some agencies want to give, because schema and llms.txt are easy to sell as technical fixes. Here is the longer version.

What is schema markup?

Schema markup (also called structured data) is a block of code on your web pages that labels information in a standard format. Instead of a search engine having to work out that “Mon to Fri 8am to 5pm” is your opening hours, schema states it directly.

For a local business, the useful types are usually:

• LocalBusiness (or a more specific type such as Bakery, Plumber or Dentist) with your name, address, phone, hours and service area
• FAQPage for a genuine FAQ section
• Product or Service where relevant
• Review data only where it follows Google’s rules, which are strict about self-serving reviews

Does schema help with Google?

Yes, in a modest and well documented way. Google’s Search Central documentation explains how structured data helps Google understand page content and can make pages eligible for certain rich results. That is useful.

But Google has also been clear in its guidance on AI features in Search that there are no special requirements or extra markup needed to appear in AI Overviews. The same fundamentals that help normal search apply. So schema supports understanding; it is not an AI ranking factor you can switch on.

Does schema help with ChatGPT, Perplexity or Copilot?

Possibly a little, indirectly. When these assistants search the web, they read your page content. Clean, consistent facts help, whether they are in visible text or in structured data. But no major AI company has said that schema markup makes its assistant more likely to recommend a business.

What we see in practice is that businesses with excellent schema and weak listings still get skipped, while businesses with no schema at all but strong reviews and correct listings get named regularly. Schema is not the deciding factor.

What is llms.txt?

llms.txt is a proposed standard, put forward in 2024, for a plain text file at the root of a website that gives AI tools a tidy summary of the site and links to its most important pages. The idea is reasonable: help language models find the useful content quickly.

The problem is adoption. There is no public confirmation that ChatGPT, Gemini, Perplexity or Copilot use llms.txt when deciding which businesses to mention. Google representatives have said publicly that Google Search does not use it. Some developer tools and documentation sites use it, which is where the idea came from, but that is a different use case from local business recommendations.

Is there any harm in adding llms.txt?

Not much, as long as it is accurate. The risk is the opposite of what people expect: an llms.txt file that is out of date or contradicts your listings could add another source of conflicting information. If you add one, keep it short, factual and in sync with your Google Business Profile.

The bigger harm is opportunity cost. Time spent perfecting a file nobody confirms is read is time not spent fixing your Bing listing or asking customers for reviews.

So what actually moves AI visibility for local businesses?

From everything we know about how assistants gather local information, the levers that matter most are:

1. Consistent listings. Google Business Profile, Bing Places, Apple Business Connect and the main directories for your trade, all with the same name, address, hours and categories. See where AI assistants get local business information.
2. Genuine reviews. A steady flow of recent reviews with real detail gives assistants something to summarise. Our guide on how many Google reviews you need covers this.
3. Clear website facts. Plain sentences that say what you do, where, and for whom, plus an FAQ that answers real questions.
4. Independent mentions. Local news, trade association listings, community directories and partner pages.

Schema sits underneath point 3 as a tidy way to express the same facts. llms.txt, at best, sits in the same place. Neither replaces points 1, 2 or 4.

What should a sensible schema setup look like?

If you are going to do it, do it properly and keep it simple:

• One LocalBusiness block (or the most specific type that fits) on the homepage or contact page
• Name, address and phone matching your Google Business Profile exactly
• Opening hours that match your listings, updated when they change
• FAQPage markup only on pages that show those FAQs visibly
• No fake or self-written review markup
• Tested with Google’s Rich Results Test after changes

That is usually a short job for whoever looks after your website. If it turns into weeks of work, something has gone wrong.

How should you think about AI search “hacks” in general?

A good rule: if a tactic is invisible to customers and promised to change what AI says about you, be sceptical. Assistants are designed to summarise what the wider web says about a business. The reliable way to change the summary is to change what the web says, through accurate listings, real customer feedback and genuine mentions.

For a fuller plan, see how to get recommended by ChatGPT, or read how we approach this on our AI search page.

If you want to know where you actually stand, our free AI Visibility Check asks the main assistants real customer questions about your business and tells you, by email, what they said and which fixes are likely to matter most. No technical jargon and no obligation.