Skip to content

LLM SEO for online stores

Short answer

LLM SEO is the technical work that lets AI assistants built on large language models find, read and cite your product pages. For a store it means allowing the right crawlers, serving key product facts in the HTML without JavaScript, structured data that matches the page, current product feeds and clear internal links. llms.txt is optional, and no step can promise a mention.

By , Founder, SourceVision. Published .

On this page
  1. What is LLM in SEO?
  2. How do LLMs find websites?
  3. How to optimize for LLMs?
  4. Do LLMs use schema markup?
  5. Is llms.txt worth it?
  6. Is SEO still relevant in 2026?
  7. Frequently asked questions
  8. Related guides
  9. Sources

LLM SEO is the technical side of being found by AI assistants. This guide is for the person who runs your store’s site: the theme, the robots.txt, the CDN, the feeds. It covers how large language models reach and read product pages, what their makers document, and the work that follows, in the order it pays off.

For the wider picture of what AI visibility means and how to measure it, start with our guide to AI visibility.

What is LLM in SEO?

An LLM, a large language model, is the system behind ChatGPT, Gemini, Claude and Perplexity’s answers. LLM SEO is the work of making your pages easy for those assistants to find, read and cite. It builds on technical SEO and adds crawlers, feeds and files.

The phrase has a second meaning in engineering, where “LLM optimization” means making a model run faster. That is not this guide. Here it means getting your products into the answers.

The names vary. Google’s optimization guide lists AEO and GEO, and adds that for Google Search, optimizing for generative AI search is “still SEO.” Our GEO vs SEO guide compares the labels, and our guide to answer engine optimization covers AEO for stores. This one stays on the technical work.

How do LLMs find websites?

Through three routes: search crawlers that index your pages ahead of time, fetchers that open a page when a shopper asks, and product data that your platform or feed sends directly. Each assistant’s maker documents its own crawler names and rules. Training crawlers are a separate setting.

The model itself does not browse. The assistant around it runs searches, fetches pages and reads product data, then the model writes the answer from what came back. Google says its AI features may run a query fan-out, several related searches at once, to gather supporting pages.

How an assistant turns a question into an answer
  1. A shopper asks

    The model itself does not browse.

  2. The assistant runs searches

    Google's AI features may run several related searches at once.

  3. It fetches pages and reads product data

    From its search index, a live fetch or a feed.

  4. The model writes the answer

    From what came back.

Assistants can only use what their crawler can fetch and read, so access comes before everything else.

Step through the four kinds of access below. Each one has a different control.

Four ways an AI assistant reaches your product data

  1. 1 Search crawlers

    OAI-SearchBot (ChatGPT search), PerplexityBot (Perplexity search), Claude-SearchBot (Claude search) and Googlebot (Google Search, including AI Overviews and AI Mode). These build the index answers draw on. OpenAI says sites that block OAI-SearchBot are not shown in ChatGPT search answers.

  2. 2 User-triggered fetchers

    ChatGPT-User, Perplexity-User and Claude-User open a page when a person asks about it. OpenAI and Perplexity say robots.txt rules may not apply to these, because a user started the request.

  3. 3 Training crawlers

    GPTBot and ClaudeBot collect content that may train future models. Each is a separate robots.txt setting, so you can block training and still allow search. Google-Extended also controls grounding in the Gemini app, not only training.

  4. 4 Catalogs and feeds

    Product data sent directly: Shopify Catalog for Shopify stores, Google Merchant Center feeds, and OpenAI's direct product feed, which merchants apply for. Shopify says its Catalog works independently of robots.txt, but Google still crawls the product pages a Merchant Center feed links to.

The sources: OpenAI’s crawler overview, Perplexity’s crawler docs, Anthropic’s crawler help and Google’s crawler list.

Two details catch technical teams out. Google says Google-Extended does not affect inclusion in Google Search, but it does govern whether Google may use your content for grounding answers in the Gemini app. And Shopify’s robots.txt help says Shopify Catalog sends product data to the agentic storefronts you have turned on, such as ChatGPT, independently of robots.txt. Blocking a crawler there only removes your open-web pages.

How to optimize for LLMs?

Start with access and readability. Allow the search crawlers in robots.txt and in your CDN or firewall, then make sure each product page states its key facts in plain HTML without needing JavaScript to run. Assistants can only use what their crawler can fetch and read.

Crawler access. OpenAI says sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers. Its search help adds that your host or CDN must also allow traffic from OpenAI’s published IP addresses. A robots.txt that allows a crawler is not enough if a firewall rule blocks it. For Google, also check Search Console’s generative AI control. Inclusion is the default, and a site excluded there is not eligible for Google’s AI features.

The crawler names to check in robots.txt, by assistant
AssistantSearch crawlerUser-triggered fetcherTraining crawler
ChatGPT (OpenAI) OAI-SearchBotChatGPT-UserGPTBot
Perplexity PerplexityBotPerplexity-UserNone named in this guide
Claude (Anthropic) Claude-SearchBotClaude-UserClaudeBot
Google GooglebotNone named in this guideGoogle-Extended, which also controls Gemini app grounding

Training crawlers are a separate robots.txt setting, so you can block training and still allow search.

CDN settings change too. Cloudflare’s AI bot settings sort AI crawlers into Search, Training and Agent, each with its own choice. On Shopify, the platform handles bot management at the network layer, and Shopify does not recommend a proxy in front of the store.

Watch for old robots.txt edits. Shopify warns that a theme with an older robots.txt.liquid template can keep rules the current default no longer has, such as Disallow: /policies/, which asks crawlers that follow robots.txt to stay out of your policy pages. Deleting the template restores the current default. Shopify also says robots.txt rules are advisory, and not every crawler follows them. Check what your robots.txt allows. The checker reads robots.txt only, so check your CDN or firewall settings separately:

AI crawler access checker

Check whether 15 AI crawlers, from GPTBot to PerplexityBot, may read your site's pages.

For example yoursite.com.

A quick check confirms you are a person, not a bot.

Descriptions checked 2026-10-13. We do not save the address or the result.

Open the full tool

Rendering. Google renders JavaScript with a headless browser, but it also says in its JavaScript basics that server-side or pre-rendering is still a great idea, because “not all bots can run JavaScript.” OpenAI, Perplexity and Anthropic do not document whether their crawlers run JavaScript. So do not rely on it.

A quick test: open a product page with JavaScript turned off in your browser, or view the raw page source. These should be in the HTML:

  • Product name, price, availability and variant options.
  • The description and the answers to common questions.
  • Shipping and return terms, or a plain link to them.
  • Links to the category and related products.

Check apps too. Reviews, size charts or FAQs that an app loads with JavaScript after the page loads may be missing from that source view.

Do LLMs use schema markup?

Some do, in some form. OpenAI says ChatGPT shopping weighs structured metadata such as price and description, and Google says structured data is not required for its AI features but still helps with rich results. Treat markup as a second copy of the page’s facts: complete, current and matching the visible text.

Google’s AI features page lists making sure structured data matches the visible text among the SEO basics for AI Overviews and AI Mode. It also says there is no special schema to add. OpenAI’s shopping help says ChatGPT considers structured metadata from first-party and third-party providers, such as price and product description.

One technical rule matters for stores. Google warns that dynamically generated Product markup can make Shopping crawls less frequent and less reliable, in its guide to JavaScript markup. Price and stock change often, so put Product markup in the server HTML. Our guide to product structured data covers the properties.

Feeds. Microsoft’s GEO playbook names product feeds and APIs as one of three inputs to AI shopping answers, next to crawled data and the live site. OpenAI’s feed spec asks for a full snapshot at least daily. Keep feed, markup and page saying the same price and stock. Our guide to AI shopping programs covers each feed.

Internal links and page experience. Google lists both among the basics for its AI features: content findable through internal links, and a good page experience. No assistant maker we read documents a separate speed threshold for AI answers. Link every product from at least one category page, and fix speed for shoppers first.

Sitemaps. Keep the XML sitemap complete and its dates honest. Google says in its sitemap guide that it uses the lastmod date only when it is consistently and verifiably accurate, and that lastmod should reflect the last significant update. A theme that stamps every page with today’s date gives Google a reason to ignore the field.

Is llms.txt worth it?

It is optional and low cost, not a ranking lever. Google says Google Search ignores llms.txt. Chrome’s Lighthouse marks the file optional for now. Shopify already serves a default /llms.txt for every store. Keep yours accurate if you have one, and spend your first hour on crawler access instead.

llms.txt is a proposal from September 2024: a Markdown file at the site root with a short summary and links to key pages. Google’s guide is plain about Google Search: it ignores llms.txt, so the file neither helps nor harms your visibility there. The Lighthouse audit marks a missing file as not applicable, because providing it is optional at the moment.

On Shopify, the agents.md docs say Shopify manages an /agents.md file for every store, and /llms.txt and /llms-full.txt mirror it by default. A custom template means you keep it current yourself. On other platforms, check what you have or write one here:

llms.txt generator and checker

Check your site's llms.txt, or write one from your own pages, for Shopify, WooCommerce, Wix and any other online store.

For example yoursite.com.

A quick check confirms you are a person, not a bot.

It reads only what your site already publishes. We do not save the address or the file.

Open the full tool

Our guide to llms.txt covers what to put in the file.

Is SEO still relevant in 2026?

Yes. Every route above starts from technical SEO: crawlable pages, server-rendered content, valid structured data, sitemaps and internal links. LLM SEO adds more crawlers to allow, feeds to keep in sync and answers to measure. The map below ties each item to our 8-step series.

Microsoft’s playbook says traditional SEO stays essential because AI systems run real-time web searches throughout the shopping journey. Here is the technical work, mapped to our 8-step series:

Technical itemSeries stepFree tool
Search crawlers allowed in robots.txt and the CDNStep 1: AI crawlersAI crawler checker (robots.txt only)
llms.txt accurate, or the platform defaultStep 2: llms.txtllms.txt generator and checker
Product markup in the server HTML, matching the pageStep 3: structured dataStructured data checker
Shipping and return terms in plain HTMLStep 4: store policiesReturn policy tool
Answers to shopper questions in the HTML, not loaded laterStep 5: product questionsProduct FAQ generator
Feeds current and matching the pageStep 6: AI shopping programsNone
Answers measured across runsStep 7: check visibilityNone
Rechecked after theme and app changesStep 8: monthly checklistAll of the above

The last two rows are where technical work meets results. A crawler log tells you a bot came. It does not tell you whether the assistant named your product, or what it said. For that you ask the shopper’s questions, several times each, on each assistant, and count.

That is the part SourceVision does. We ask ChatGPT, Gemini and Perplexity the questions your shoppers ask, many times over, and report how often your products appear, with the range across runs. Then we propose fixes for each product page, drafted only from facts you confirm, and nothing is written without your approval. See how it works.

SourceVision is in an invite-only beta.

Request an invite

None of the work above can promise a mention. OpenAI says plainly that it does not promise placement in ChatGPT search. What it does is remove the technical reasons an assistant has to skip you, and you can check each one.

Frequently asked questions

What is SEO for LLMs called?

It goes by several names: LLM SEO, LLM optimization, GEO for generative engine optimization, and AEO for answer engine optimization. Google lists AEO and GEO as terms for work focused on visibility in AI search, and says that for its own AI features the work is still SEO.

Is SEO dead now with AI?

No. Google says its AI Overviews and AI Mode need no special optimization beyond being indexed and eligible for a snippet, as long as the site is not excluded with Search Console's generative AI control. Microsoft's GEO playbook says traditional SEO remains essential because AI systems run live web searches during the shopping journey.

Can ChatGPT do SEO?

ChatGPT can help draft titles, descriptions and answers, but it cannot change how your site is crawled or indexed. The technical work, such as crawler access, rendering and structured data, happens on your site and in your platform settings. Check every fact it drafts against your product before you publish it.

What is LLM optimization?

In marketing, LLM optimization means making your content and products easy for assistants built on large language models to find and use in answers. The same phrase also describes making the models themselves run faster, which is a separate engineering topic. For a store, it is the first meaning.

How do you optimize for LLMs?

Let the assistants' search crawlers reach your pages, serve key product facts in the HTML, keep structured data and feeds matching the page, and link every product from a category page. Then ask your shoppers' questions on each assistant several times and count how often your products appear.

Does llms.txt work?

There is little evidence that it changes AI answers. Google says Google Search ignores llms.txt, and none of the OpenAI, Perplexity or Anthropic crawler pages we read says their crawlers use it. It does no harm, and Shopify already serves a default file for every store.

  • GTIN, UPC and barcodes for online sellers

    What GTIN, UPC and EAN codes are, where to get them, which ones Shopify, Google, ChatGPT and Amazon ask for, and what to do when a product has none.

  • Product descriptions for AI shopping

    How to write product descriptions that shoppers, search engines and AI assistants read correctly: structure, facts with units, feeds and what to leave out.

  • Product titles for AI shopping

    Write product titles for AI shopping, Google Shopping and marketplaces: what to include, in what order, each channel's length limit and what to leave out.

Sources

  1. Overview of OpenAI crawlers, OpenAI, accessed .
  2. Searching the web with ChatGPT, OpenAI Help Center, accessed .
  3. Shopping with ChatGPT Search, OpenAI Help Center, accessed .
  4. File upload overview (product feeds), OpenAI, accessed .
  5. Perplexity crawlers, Perplexity, accessed .
  6. Does Anthropic crawl data from the web, and how can site owners block the crawler?, Claude Help Center, accessed .
  7. List of Google's common crawlers, Google, accessed .
  8. AI features and your website, Google Search Central, accessed .
  9. Guide to optimizing for generative AI features on Google Search, Google Search Central, accessed .
  10. Understand the JavaScript SEO basics, Google Search Central, accessed .
  11. Generate structured data with JavaScript, Google Search Central, accessed .
  12. From discovery to influence: a guide to GEO (playbook PDF), Microsoft Advertising, accessed .
  13. Editing robots.txt.liquid, Shopify Help Center, accessed .
  14. agents.md.liquid, Shopify.dev, accessed .
  15. Block AI Bots, Cloudflare Docs, accessed .
  16. The /llms.txt file, llmstxt.org (Jeremy Howard), accessed .
  17. llms.txt (Lighthouse agentic browsing audit), Chrome for Developers (Lighthouse), accessed .
  18. Search generative AI control, Search Console Help, accessed .
  19. How to fix: Unable to check product pages, Google Merchant Center Help, accessed .
  20. Build and submit a sitemap, Google Search Central, accessed .

The beta is invite only for now.

Free during the founding beta. Founding partners get 50% off list price after the beta.