Short answer: Whether ChatGPT, Claude, Gemini, and Perplexity recommend your store comes down to eight fixable things, grouped by a simple question — can they read your store (product schema, a reachable catalog feed, barcodes), can they trust it (published policies, findable shipping), and can they reach it (AI crawlers allowed, an llms.txt, a conversational endpoint)? Most are free to fix, and you can check all eight in about two minutes.
More shoppers are opening ChatGPT or Perplexity instead of a search engine and asking what to buy. The assistant replies with a short list of stores — and if yours isn't on it, you never got the chance to compete. The uncomfortable part is that being left off usually has nothing to do with how good your products are. It comes down to a handful of unglamorous, mostly-technical things that decide whether an assistant can read, trust, and reach your store at all.
Here's the whole checklist, then the reason each item matters and how to fix it.
The checklist
- Product schema on every product page
- A catalog feed machines can actually reach
- Barcodes (GTIN/UPC) across your feed
- Store policies published — returns, privacy, terms
- Shipping answerable right on the product page
- AI crawlers allowed in your robots.txt
- An llms.txt at your domain root
- A conversational endpoint an agent can actually use
Three questions cover all eight. Work through them in order.
Can AI assistants read your store?
Before an assistant can recommend you, it has to understand what you sell — in a form built for machines, not just human eyes.
Product schema. Structured data (JSON-LD) on each product page spells out price, availability, and images in a format assistants and shopping agents parse reliably. Most platforms support it; the catch is that it only works when your product fields are actually filled in. We went deeper on this in why AI recommends some stores and skips yours.
A reachable catalog feed. Shopping surfaces ingest your catalog through a feed. If it's missing, gated, or malformed, your products simply aren't in the pool the assistant draws from. This is a yes/no gate: either the feed answers when a machine asks, or it doesn't.
Barcodes (GTIN/UPC). This is the single most common blocker we see in store audits — feeds with barcodes on almost none of their products, sometimes none at all (we shared the numbers in three findings from 25 store audits). The barcode is the join key that attaches your listing to the canonical product card where shopping AIs compare sellers. Without it you can still be found by name, but in the anonymous "best X" searches where most demand starts, you float unattached while the card lists your competitors. The fix is an afternoon in your admin — here's how.
Can they trust your store?
Reading you isn't enough. Assistants — especially the ones with shopping surfaces — won't confidently send a shopper to a store whose basics they can't verify.
Published policies. Clear returns, privacy, and terms pages aren't just good manners; some shopping surfaces require them before they'll transact. If yours are missing or unreachable, that's a hard stop the assistant won't explain to the shopper — it'll just pick someone else.
Shipping the agent can find. "How much is shipping?" is one of the questions buyer agents ask most, and one they most often can't answer — because the info lives three clicks deep in a policy page, not near the price where the decision gets made. A one-line shipping summary on the product page closes that gap. More in store policies and shipping.
Can they reach your store?
Finally, the assistant has to be allowed to read you — and ideally handed your story in your own words.
AI crawlers allowed. Check your robots.txt. If it blocks the AI agents by name — GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, and their kin — you've locked the front door to the exact visitors you want. Welcoming them is a one-line change.
An llms.txt. A short plain-text file at your domain root that tells assistants who you are, what you sell, and which pages matter — in your own words, so they don't reconstruct your story from forum posts and third-party guesses. It's an emerging convention worth getting ahead of. What to put in it, and why in where AI agents read about your brand.
A conversational endpoint. The hardest and most valuable one. Static pages answer the questions you anticipated; they can't handle "which of these fits an overpronator with wide feet?" — the back-and-forth that turns a browser into a buyer. Today, almost no independent store has a door an outside agent can knock on to ask. It's the gap we started sqrly to close, and it barely exists in the wild yet.
How do I check — and fix — all eight?
You don't have to audit these by hand. The sqrly Pulse — free snapshot runs the whole list against your store and shows you, in about two minutes, where you stand on each — cold-tested the way a real shopper's assistant would encounter you. It's free, works for any online store, reads your catalog directly on Shopify and WooCommerce, and needs nothing but your URL: assessment.sqrly.ai/check.
What changed recently is what happens after the diagnosis. Pulse no longer just flags a gap and wishes you luck — it hands you the fix. For some items that's a ready-to-use draft you review and publish: it writes your llms.txt from your own store's data, and when an assistant states something false about you — the wrong shipping, hours, or policy — its correction kit drafts the corrected page copy and points you at the exact page that needs it. For the rest, you get store-specific, step-by-step instructions for your platform, not generic advice. Nothing is ever published to your store without you.
If you want the list watched over time — re-checked as you fix things, as assistants re-crawl, and as competitors move — that's what sqrly Pulse does for $9.99 a month.
But the checklist is the point, not the tool. Every item here is something you can verify, and mostly fix, yourself. Being recommended by AI assistants isn't luck or budget — it's whether the new kind of visitor can read you, trust you, and reach you. Eight boxes. Start ticking.
sqrly helps online retailers show up — knowledgeably — for every shopper and every AI assistant. Join the waitlist for the sqrly agent, or run your free snapshot.