llms.txt and Structured Data for MSPs: What Works

For a managed service provider (MSP), llms.txt is cheap hygiene, not a lever. It is a proposed plain-text map of your best pages for artificial intelligence (AI) tools, and Google's written guidance says Google Search ignores it. Structured data does more of the useful work: schema.org markup in JSON-LD (JavaScript Object Notation for Linked Data) states who you are, where you work and what you sell in a fixed format. That can help Google distinguish your organization from others and gives other systems consistent business facts. Add both in an afternoon, then put your real effort into pages and proof.

This page is the practical version for an MSP site. For the strategy behind AI search, start with NUOPTIMA's generative engine optimization (GEO) for MSPs page.

What llms.txt Is, and What It Is Not

llms.txt is a proposal published by Jeremy Howard in September 2024 at llmstxt.org. The idea: put a Markdown file at the root of your site that gives a language model a short, clean summary of who you are and links to the pages worth reading, without the menus, scripts and cookie banners a crawler has to dig through. The format is simple. One top-level heading with your name is the only required part. A short quoted summary follows, then sections that list links, plus an optional section an agent can skip when it is short on space.

It is not a standard, a sitemap replacement or a ranking file. Nothing in the proposal forces any AI company to read it.

What Google and the AI Assistants Have Said

Google has been unusually direct. Its guide to optimizing for generative AI features lists llms.txt and other special markup under things you can ignore, and says adding these files "will neither harm nor help" your visibility in Google Search, because Google Search ignores them. Its AI features page says the same in other words: no new machine-readable files and no special schema are needed to appear in AI Overviews or AI Mode.

The vendor documentation cited here does not establish llms.txt as a ranking or citation factor for any assistant. Some agents may fetch it, and it costs nothing to serve, so the honest position is: add it, keep it accurate, and expect nothing measurable from it on its own. If a vendor sells llms.txt as the reason you will start appearing in ChatGPT, walk away.

What a Good MSP llms.txt Looks Like

Keep it short and factual. Write the summary the way you would want an assistant to describe you: services, the industries you serve, the cities you cover and how long you have been in business. Link the pages a buyer actually needs, not every blog post. A workable outline for a fictional Dallas provider of managed information technology (IT):

# Example IT Partners

> Managed IT and cybersecurity for law firms and manufacturers
> in Dallas and Fort Worth, Texas. Founded [year]. Office: [street
> address], Dallas, TX. Serves offices of [x] to [y] staff.

## Services
- [Managed IT](https://www.example.com/managed-it): flat monthly fee, help desk and on-site support
- [Cybersecurity](https://www.example.com/cybersecurity): monitoring, backup, incident response
- [Co-managed IT](https://www.example.com/co-managed-it): support for an in-house IT team

## Industries
- [Law firms](https://www.example.com/industries/law-firms)
- [Manufacturing](https://www.example.com/industries/manufacturing)

## Locations
- [Dallas](https://www.example.com/locations/dallas)
- [Fort Worth](https://www.example.com/locations/fort-worth)

## Proof and pricing
- [Case studies](https://www.example.com/case-studies)
- [Pricing](https://www.example.com/pricing)

## Optional
- [Blog](https://www.example.com/blog)

Save it as a plain text file at yoursite.com/llms.txt and update it whenever you add a service, an office or a location page. A stale file that lists a closed office is worse than no file.

How NUOPTIMA Runs Its Own

NUOPTIMA serves an llms.txt file of its own. Each site build regenerates its page links from the same content records that produce the pages, so new guides and case studies appear without a manual edit. The summary and contact details at the top are written by hand, which makes them the part most likely to drift out of date, so put that part of your own file on a review calendar. It opens with a summary and a list of key pages, then contact details, then sections for our MSP services, industries, guides and case studies. A longer companion file, llms-full.txt, indexes every page, and each page also has a clean Markdown version for agents. We publish it because it costs us little to maintain, not because we can point to a ranking or citation it produced.

The Structured Data Types That Matter for an MSP

Structured data earns its place for a different reason. Google says Organization markup can help it disambiguate your organization from others, and a consistent set of facts gives every machine reading your site the same answer about who you are.

The entity work behind the first two rows is covered in NUOPTIMA's guide to knowledge graph search engine optimization (SEO). If you run as a service-area business, decide how you present your address first, using the guide to service area vs office address, then make the markup match.

What Google Still Shows, and the Review Star Rule

Two changes catch MSPs out. First, frequently asked questions (FAQ) rich results are gone. Google limited them to well-known government and health sites in 2023, then stopped showing them for everyone from May 7, 2026, and removed the documentation in June 2026. FAQPage remains a valid schema.org type, so keep it where it describes a real questions section, but do not expect dropdowns under your listing.

Second, review stars. Google's review snippet rules say that when a business controls the reviews about itself, its LocalBusiness or Organization pages are not eligible for star results, including reviews pulled in through a third-party widget. Marking up your own testimonials with aggregateRating on your home page will not earn stars. Google says self-serving review markup on its own does not lead to a manual action, but fake reviews, or ratings that visitors cannot see on the page, still break its structured data guidelines. Let your Google Business Profile carry your stars. The Google Business Profile guide for MSPs covers that side.

No markup unlocks Google's AI features. Its value is plainer: consistent, machine-readable facts, and eligibility for the rich results that still exist.

A ProfessionalService Example

Here is a short block for a fictional Dallas MSP. It goes on the home page inside a script tag of type application/ld+json, and the facts must match the visible page, your Google Business Profile and your directory listings exactly.

{
  "@context": "https://schema.org",
  "@type": "ProfessionalService",
  "@id": "https://www.example.com/#business",
  "name": "Example IT Partners",
  "url": "https://www.example.com/",
  "logo": "https://www.example.com/logo.png",
  "description": "Managed IT and cybersecurity for law firms and manufacturers in Dallas and Fort Worth.",
  "telephone": "+1-214-555-0100",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "100 Main St, Suite 200",
    "addressLocality": "Dallas",
    "addressRegion": "TX",
    "postalCode": "75201",
    "addressCountry": "US"
  },
  "areaServed": [
    { "@type": "City", "name": "Dallas" },
    { "@type": "City", "name": "Fort Worth" }
  ],
  "sameAs": [
    "https://www.linkedin.com/company/example-it-partners"
  ],
  "makesOffer": {
    "@type": "Offer",
    "itemOffered": {
      "@type": "Service",
      "name": "Co-managed IT",
      "serviceType": "Managed IT services"
    }
  }
}

One caveat. The schema.org page for ProfessionalService carries a note that the general type was deprecated because people confused it with Service. It still validates and is widely used. If you want zero ambiguity, use LocalBusiness instead. Pick one and use it everywhere.

Why Entity Clarity Pays Off

When NUOPTIMA audited Chrysalis Tech Group, a Houston-area MSP, in September 2026, some AI assistants asked about Chrysalis by name described a different company. The rebuilt site, live in October 2026, has structured data on every page naming both offices, plus a company facts page also published as plain-text files for AI assistants. It is too early to report results from it. The point is the diagnosis: when several companies share your name, consistent markup and facts give machines a way to keep them apart.

Test It Before You Trust It

Run every template through two free tools. Google's Rich Results Test shows whether a page is eligible for the rich results Google supports and flags errors in those types. Local business markup is one of them, and Google asks for the most specific LocalBusiness subtype, so run your LocalBusiness or ProfessionalService block through it. The Schema Markup Validator checks any schema.org type, including Service, which is not one of Google's rich result types. Recheck after every theme, plugin or site update, because markup breaks quietly. Our MSP SEO audit checklist includes these checks with the rest of the technical review.

Where This Fits

Both files are a small part of being recommended. The bigger levers are clear service and location pages, proof, reviews and mentions on sites buyers trust, which is what the guide to getting your MSP recommended by ChatGPT walks through. If you want to know whether your markup, crawl access and entity facts are holding you back, start with a free audit of your site and market.

Sources

• https://llmstxt.org/, accessed October 25, 2026.
• https://developers.google.com/search/docs/fundamentals/ai-optimization-guide, accessed October 25, 2026.
• https://developers.google.com/search/docs/appearance/ai-features, accessed October 25, 2026.
• https://developers.google.com/search/updates, accessed October 25, 2026.
• https://developers.google.com/search/docs/appearance/structured-data/organization, accessed October 25, 2026.
• https://developers.google.com/search/docs/appearance/structured-data/local-business, accessed October 25, 2026.
• https://developers.google.com/search/docs/appearance/structured-data/review-snippet, accessed October 25, 2026.
• https://schema.org/ProfessionalService, accessed October 25, 2026.
• https://search.google.com/test/rich-results, accessed October 25, 2026.
• https://validator.schema.org/, accessed October 25, 2026.
• https://nuoptima.com/case/chrysalis-msp-case-study, accessed October 25, 2026.
• https://developers.google.com/search/blog/2019/09/making-review-rich-results-more-helpful, accessed October 25, 2026.
• https://developers.google.com/search/docs/appearance/structured-data/sd-policies, accessed October 25, 2026.