Can an AI agent answer your store's shipping question?

Ask an AI shopping agent — a browsing assistant, a support bot, a comparison tool — "when will my order arrive" or "can I pay cash on delivery" about almost any independent online store, and there is a real chance it cannot answer. Not because the agent is bad. Because the answer was never written down somewhere the agent could find it.

We tested this directly. In July 2026 we ran a full crawl-and-audit against four real, independent Shopify stores (not a demo, not a sandbox — live stores we found through their own public product feeds). Each store was scored on ten checks that matter to an AI shopping agent: product titles, descriptions, variants, category guidance, contact info, structured data, returns, shipping, payment availability, and FAQ. The mean score was 15.8 out of 20.

The breakdown is the interesting part. Product titles, descriptions, variants, category guidance, contact information, and structured data (JSON-LD) all averaged a perfect 2.00 out of 2 across the four stores. Modern Shopify themes and apps have quietly solved that problem. Returns policy averaged 1.75. Shipping averaged 1.25.

Payment availability averaged 0.50 out of 2. FAQ content averaged 0.25 out of 2.

That is the gap. Not the product page — the two pages a buyer needs before they trust the product page enough to check out. And it is not just a handful of edge cases. In a separate, wider pass we reviewed ten independent D2C stores across the US and India for exactly this kind of gap, looking only at what was actually reachable from each store's own navigation and footer, with no guessing and no template inference.

Nine of those ten stores had a shipping-policy page that was either missing entirely, returned a 404, was mislabeled and pointed at the wrong document (in one case, both the "shipping policy" and the "returns policy" links in the footer resolved to the same privacy-policy page), or contained an unfilled template placeholder in place of a real delivery estimate. Several also had no findable FAQ page at all — one store's own "FAQs" link in the footer simply pointed back to its homepage.

None of this is a criticism of any one store. It is a pattern, and patterns are useful because they are fixable in a predictable way. A support agent, human or AI, cannot answer a question the site never actually answered. An AI shopping agent reading a broken or placeholder shipping page does not know it is broken — it will quote whatever text is there, including the placeholder, as if it were a real policy.

Why this keeps happening on well-built stores

The stores in our sample were not badly built. They ran on Shopify, had clean themes, and in most cases had already put real effort into product photography and descriptions. Policy pages get treated as a one-time setup step, filled in from a template during store launch and then never revisited. Product pages get edited constantly because they drive sales directly. Policy pages don't — until an agent or a frustrated customer hits the gap.

The fix is not a redesign. It is usually a few hours of finding every place a policy question could be asked, writing (or rewriting) the real answer once, in one place, and making sure every link that claims to point at that answer actually does.

What to check on your own store

Three checks take under ten minutes and catch most of what we found: open your shipping-policy URL directly and read it as if you were a first-time buyer, not a browsing bar checking it loaded. Click every "FAQ," "shipping," and "returns" link in your footer and confirm each lands on a different, real page. And ask, out loud, "does this page say what happens if I want to pay cash on delivery, and how would an AI agent reading this page know that." If the answer is no, that is the highest-leverage place to start.