Shopify is 80.7% agent-ready. Product schema is the real gate

Every agency serving Shopify merchants is fielding some version of the same question this quarter: is my client's store actually ready for an AI shopping agent to read it, trust it, and buy from it. Most agencies answer by feel, because nobody has had a number to point to. We built the Vergio Agentic Commerce Index so the answer has one, and we published it this month on our agentic commerce data page.

The reassuring 80.7% Agent-Ready score is mostly a llms.txt effect. The facet that actually decides whether a store passes is Product schema, not the one everyone assumes.

The headline, and what it is built from

The VACI stands at 70.7 this month, and 80.7% of the Shopify stores in our continuously monitored panel now qualify as Agent-Ready. We define Agent-Ready plainly: a store does not block AI shopping crawlers in robots.txt, it publishes machine-readable Product schema, and it offers either an llms.txt file or FAQ schema. Clear the crawler-blocking check, clear Product schema, and clear one of the last two, and a store passes.

Read that definition again and one thing stands out. Three separate conditions have to hold, and they do not hold with equal difficulty. Being Agent-Ready is not a matter of shipping one flashy feature; it is a matter of which of these three gates is still shut.

Two of the four gates are already open

Start with crawler blocking. At 0.2%, almost no monitored store is turning away AI shopping bots in its robots.txt. That gate is functionally open across the panel, and it will not be the reason a store fails the Agent-Ready check.

Now look at the "or" clause: llms.txt or FAQ schema. FAQ schema is rare at 4.3%, the kind of number that would normally read as a crisis. But it barely matters here, because llms.txt adoption is already at 97.5%. Almost every store satisfies that half of the definition through llms.txt alone, whether or not it ever adds FAQ schema. An agency that spends its energy pushing a client toward FAQ schema, thinking it closes the Agent-Ready gap, is fixing something that is not broken. llms.txt already closed it.

The gate that is actually deciding the outcome

That leaves Product schema, sitting at 81.2%. Because the other three conditions are close to universal or close to solved through llms.txt, Product schema is doing almost all of the discriminating work in whether a store lands on the Agent-Ready side of the line or not. A store that blocks nothing and serves llms.txt, but never marked up its product pages with structured data, is the profile actually failing this check today, far more often than a store caught out by FAQ schema.

That is the useful, checkable finding here. If a client's store is not Agent-Ready, the first thing worth pulling up is not its FAQ page. It is whether its product pages carry Product schema at all, since that is the condition still open at scale across the panel.

Why this matters faster than it looks

This is not a distant concern. Shopify already made the plumbing for agent-driven shopping the default on every store, part of the same Summer '26 push we covered when it shipped in June: a catalog an agent can read and act on without a custom integration. The mechanism is on. What varies store to store is whether the catalog underneath it hands an agent anything worth reading, and Product schema is the single largest reason that varies.

For an agency, that reframes a piece of technical SEO work most merchants have ignored for years into something with a live, dated reason to fix it now. A missing reviews app costs a store slow, compounding trust. A missing Product schema costs a store something faster: an AI shopping agent that cannot parse what it sells, at the exact moment agents are starting to do the shopping.

What to check this week

Pull the client and prospect list you already use for the usual app-stack audit, and add one line to it: does the store's product pages carry structured Product schema an agent can read. It is a five-minute check per store once you know to look, and most agencies have never looked, because until this index existed there was no clean way to say how common the gap was.

Run the free store check on a client or a prospect and it reads exactly this signal alongside the rest of the store's public setup, no manual crawling required. The product surfaces the same read across every store on your list, so you are not doing this one storefront at a time. Skip the FAQ page. Go straight to whether the product pages carry schema an agent can actually parse, because that single fact is deciding more Agent-Ready outcomes than the other three conditions combined.