Schema markup and llms.txt are the two technical fixes most often sold as AI SEO: the work of getting a site named and cited in AI answers, also called generative engine optimisation (GEO). Both are cheap to add. Neither has evidence that it earns AI citations. This guide sets out what the tests found, and the technical basics that do matter.

Does schema markup help ChatGPT or AI Overviews cite you?

Schema markup is structured data, usually written as JSON-LD, that labels what a page is about: an organisation, a product, an article, a set of questions and answers.

The clearest test so far comes from Ahrefs. It tracked 1,885 pages that added JSON-LD between August 2025 and March 2026 and compared them with about 4,000 similar pages that didn't:

  • Google AI Overviews: citations fell by about 4.6% relative to the comparison pages, a small decline rather than a gain. Ahrefs notes the fall may not have been caused by the schema itself.
  • Google AI Mode and ChatGPT: small rises (2.4% and 2.2%) that could not be told apart from zero.

The test covered pages that AI tools were already citing heavily, so it doesn't prove schema can never matter for a page nobody cites yet. It does mean nobody can honestly promise that adding schema will get you cited.

Google says the same about its own AI features. Its guidance for site owners states that there is no special schema.org structured data you need to add to appear in AI Overviews or AI Mode.

Is schema markup still worth adding?

For ordinary search, sometimes. Google uses structured data to show rich results, such as review stars or recipe details, for the page types it supports. If your pages qualify for a rich result, the markup can be worth it for that reason alone.

What it shouldn't be is your AI search plan. This site carries basic schema markup because it costs nothing to maintain. We don't expect it to earn citations.

What is llms.txt?

llms.txt is a proposal published by Jeremy Howard in September 2024: a Markdown file at the root of a site, such as example.com/llms.txt, that gives AI tools a short description of the site and links to its most useful pages. It's a convention, not a standard the AI companies have committed to read.

Does an llms.txt file help you get cited?

No clear effect has been found. Search Engine Journal reported SE Ranking's analysis of about 300,000 domains: about 10% had an llms.txt file, and having one didn't appear to change how often a domain was cited in AI answers. That analysis dates from November 2025; we know of no later study that contradicts it.

A July 2026 review of 45 studies on generative engine optimisation reached a broader version of the same conclusion: no technique it reviewed showed a stable, long-term effect across AI platforms on how often a site is found.

Do I need an llms.txt file for AI search?

No. If your site already has one, it does no harm, and this site publishes one because it took a few minutes. If you don't have one, adding it is not a priority, and nobody should charge you much for it.

Which technical basics do matter for AI search?

AI assistants can only cite pages they can reach and find. These checks matter more than any markup:

  • Let the AI search crawlers in. Check that robots.txt doesn't block them. OpenAI's crawler documentation names OAI-SearchBot as the crawler for ChatGPT search and recommends allowing it; Anthropic and Perplexity run their own. A firewall or "bot protection" setting can block them even when robots.txt allows them.
  • Be findable in more than one search engine. Each assistant searches a different index. In 1,481 matched searches, 87.8% of the pages Claude retrieved were in Brave's top ten results, against 45.6% in Google's. Submit your sitemap to Google Search Console and Bing Webmaster Tools. Brave has no sitemap route, but its crawler help page links to a form for submitting individual pages.
  • Put the words in the HTML. If the main text of a page only appears after JavaScript runs, some crawlers may see an empty page. Serving the text in the HTML itself is the simplest way to make sure any crawler can read it.
  • One clear page per question. A page that answers one narrow question directly is easier for an assistant to quote than a long page that covers everything. Our guide to writing content AI assistants cite covers this.

If you're not sure whether AI assistants already mention your brand, a basic AI visibility audit shows where you stand before you change anything.

What we saw when we asked

On 13 November 2026 we put the questions below to Claude (claude-sonnet-5-5, with web search on), 3 times each, as part of our own measurement of this site.

  • “Does schema markup help my website get cited by AI chatbots?”
  • “Do I need an llms.txt file for AI search?”
  • 6 of 6answers searched the web
  • 39sources cited, from 19 different websites
  • 0citations of lpageo.com

Counts from a single day, before this page existed. Answers change from run to run, so one day is a snapshot, not a trend. We'll keep measuring the same questions.

Sources

  1. We tracked 1,885 pages adding schema. AI citations barely moved., Ahrefs (Louise Linehan), 11 May 2026.
  2. AI features and your website, Google Search Central.
  3. Intro to how structured data markup works, Google Search Central.
  4. LLMs.txt shows no clear effect on AI citations, based on 300k domains, Search Engine Journal (Matt G. Southern), reporting SE Ranking data, 20 November 2025.
  5. The /llms.txt file, llmstxt.org (Jeremy Howard), 3 September 2024.
  6. Optimizing visibility in generative engines: a critical survey of generative engine optimization (2023–2026), Olivier Martinez, arXiv 2607.14035, 15 July 2026.
  7. Overview of OpenAI crawlers, OpenAI.
  8. Nearly nine in ten URLs Claude retrieves are in Brave's top 10, Complete SEO (Ryan Howard), 1 September 2026.
  9. Brave Search crawler, Brave.