How AI shopping assistants find and recommend your store

More shoppers now open ChatGPT, Perplexity or Google's AI Overviews before they open a store. If those assistants can't read and understand your catalog, they can't recommend you - and you never even see the customer you lost. Here is how they actually "see" a store, and what to do about it.

How an AI assistant reads a store

AI shopping assistants don't browse like a person. They crawl and parse. They lean on a handful of machine-readable signals to understand what you sell, whether it's in stock, and whether you're trustworthy enough to recommend. Get those signals right and you become quotable; get them wrong and you're invisible.

The three things they rely on

• Crawler access. Assistants use bots - GPTBot, PerplexityBot, ClaudeBot, Google-Extended and others. If your robots.txt blocks them (many stores do, unknowingly), you've opted out of being recommended.
• Structured data. Product schema (schema.org Product JSON-LD), OpenGraph tags and clear on-page text tell an assistant the name, price, availability and identifiers of what it's looking at. Without them, it has to guess - and often skips you.
• A map of your catalog. A sitemap, a product feed, and increasingly an llms.txt file help assistants find and understand the full catalog rather than a random page.

Are you accidentally blocking them?

Start with robots.txt. Open yourstore.com/robots.txt and look for Disallow rules that hit the AI crawlers above, or a blanket block. Plenty of stores inherited an aggressive robots.txt from a plugin or a well-meaning "block scrapers" tweak - and quietly excluded themselves from the assistants sending buyers today.

Make your catalog machine-readable

Every product page should expose complete Product structured data: a descriptive name, a real description, price and currency, availability, images, brand, and product identifiers (GTIN/MPN). This is the single highest-leverage thing you can do - it's the difference between an assistant confidently recommending your product and vaguely mentioning "a similar item."

Give them a map

Keep an up-to-date XML sitemap, publish a clean product feed (Google Merchant or the emerging agent feed formats), and consider an llms.txt that points assistants at the parts of your site that matter. These are cheap to maintain and directly improve how completely you're understood.

A two-minute self-check

You don't have to guess. Our free AI Store Readiness Checker scans your robots.txt, structured data, sitemap and llms.txt and scores how ready you are; the AI Product Checker grades a single product page the way an assistant reads it. Both are free and need no signup - a fast way to find the gaps before you fix them.