Does Your Store Show Up in AI Shopping Answers? Check These Five Areas

"I don't even know if we show up." That is how most ecommerce leads describe their position when the conversation turns to ChatGPT, Google AI Mode, Copilot and Perplexity. The follow-up question is usually sharper: why do we appear as a link but never as a product card? Both questions have the same answer. It depends on what the assistant can read, verify and trust about your products.

This guide walks through five areas that decide whether a store's products show up in AI shopping answers. Shopify released a free readiness scorer in late April 2026 that runs 31 checks across five categories for any store URL, and the categories it chose line up with the checks below. The areas apply whether you run on Shopify, BigCommerce, Adobe Commerce or a custom stack.

How AI shopping answers choose products

A traditional search returns a list and the shopper picks. An AI assistant works differently. Google describes AI Mode as using query fan-out, breaking a question into subtopics and issuing many searches at once, then assembling an answer from what it retrieves. A request like "running shoes for flat feet under $150 with a wide toe box" becomes several narrower questions about arch support, width, price and availability.

If your product page and feed state the facts those narrower questions ask for, you can be matched. If they do not, a competitor whose page states them is easier to recommend. Fewer fields means fewer sub-queries you can match. That is the mechanism behind each of the five areas.

Area 1: Crawler access. Can the assistant reach your pages at all?

Crawler access covers whether the systems behind each assistant can fetch your product pages, whether robots.txt permits them, and whether your sitemap exposes product URLs. Some retailers blocked GPTBot, ClaudeBot and PerplexityBot in 2024 and 2025 to keep their content out of model training. Several of those rules also block the separate search crawlers that feed shopping answers, so the decision deserves a second look.

The providers document different agents for different jobs. OpenAI's crawler documentation separates OAI-SearchBot for ChatGPT search from GPTBot for training, and Google's guidance for AI features in Search says Googlebot and normal Search controls govern AI Overviews and AI Mode. Google also says no special AI file or markup is required for its AI features, so an llms.txt file is a convenience for some crawlers, not a fix for a verified block.

What to check:

• robots.txt allows the search crawlers you want: OAI-SearchBot, Googlebot, PerplexityBot, and whichever others you decide to admit. Low effort.
• Your sitemap includes product URLs with lastmod timestamps. Low effort.
• Your server returns the same content to bots as to shoppers. Bot challenges and WAF rules that serve a blank page to crawlers remove you from consideration. Medium effort.
• Optionally, an llms.txt file at /llms.txt summarizing the site, key products and policy URLs. Low effort.

We cover the full access audit, including how to tell a search crawler from a training crawler, in are your product pages accessible to AI search.

Area 2: Product data. Can the assistant understand your products?

Product data is the layer an assistant reads to match a product against a shopper's question: JSON-LD Product markup, Offer markup with price and availability, AggregateRating and Review markup, brand and GTIN identifiers, and the attributes written into the page itself. Google's Product structured data documentation lists the properties it reads for product results. The Shopify scorer flags missing JSON-LD as a common gap.

The attributes matter as much as the markup. A page that states material, dimensions, weight, compatibility, certifications, warranty and intended use gives the assistant facts to match against the narrower questions it generates. A page with a title, a price and a two-line description leaves the assistant guessing, and guessing is where wrong answers and omissions come from.

What to ship:

• Product, Offer, AggregateRating and Review JSON-LD on every product page. Validate with Google's Rich Results Test.
• Brand, GTIN, MPN and variant identifiers populated. These are the keys assistants use to match the same product across retailers.
• The attributes shoppers actually filter on, written in the page copy in the words shoppers use, not internal SKU language. This is the core of the product content work.
• The same price, availability and attributes in your Merchant Center feed as on the page. Drift between feed and page is one of the most common reasons a product is left out or misdescribed.

Area 3: Checkout and policy data the assistant can verify

Assistants recommend, and the shopper buys on your site. OpenAI removed native checkout from ChatGPT in March 2026, as Forrester's analysis of the pullback from native checkout describes. The practical consequence is that this area is about clean redirects, accurate availability, prices the assistant can confirm, and shipping and return terms it can read.

What to check:

• Inventory in your feed matches the page and the warehouse. A recommended product that turns out to be out of stock is a wasted recommendation and a bad experience for the shopper.
• Shipping and return policies exposed in structured data, not only on a policy page. The Shopify scorer flags shipping policies that exist on the site but are not exposed in structured data.
• Referral parameters survive your checkout. Assistants pass source and UTM parameters, and checkouts that strip them lose the attribution you will need later to show what worked.

Area 4: Trust signals. Will the assistant prefer you over a brand it already knows?

Trust signals are what an assistant uses to choose between roughly equivalent products: reviews it can read in markup rather than only in rendered HTML, a consistent brand name across your site, your feed and your public profiles, HTTPS, and organization details it can confirm. Larger retailers have years of this data marked up. A brand with thousands of reviews on its own site but no Review schema gives the assistant less to work with.

Quick wins:

• AggregateRating and Review JSON-LD on every product page, sourced from the reviews you already collect.
• Organization schema with sameAs links to the profiles you control.
• One brand name, spelled the same way, across product pages, schema, Google Business Profile and Merchant Center.

Area 5: Freshness. Does the store look like a live business?

Freshness is the most overlooked area. It covers whether prices and availability update when they change, whether shipping times appear on the product page rather than only at checkout, and whether return terms are stated plainly. When two products are otherwise equal, the one with current data and clear terms is the easier recommendation.

Where to look:

• Publish ship time on the product page. Shipping promises buried in checkout are rarely read.
• State the return policy in explicit terms: window in days, free or paid, restocking fee or not.
• Match your feed refresh to how often prices and stock actually change. Daily repricing with a weekly feed sync guarantees stale answers.

Which areas are gates and which are tiebreakers

The five areas do not carry equal weight. Access and product data are gates. Fail either one and you are not in the consideration set or cannot be matched to the shopper's question. Trust signals and freshness are tiebreakers that decide between products that both pass the gates. Checkout and policy data is the closer: get it wrong and you win the recommendation, then lose the shopper on your own site.

• Crawler access: gate. If you fail it, you are not in the consideration set.
• Product data: gate. If you fail it, you cannot be matched to the shopper's question.
• Trust signals: tiebreaker. If you fail it, you lose to a more credible looking competitor.
• Freshness: tiebreaker. If you fail it, you lose to a more reliable looking competitor.
• Checkout and policy data: closer. If you fail it, you get recommended and then lose the shopper at checkout.

The free Paz AI readiness report shows which of these five areas your products currently miss, product by product. Shopify's scorer answers what is broken at the store level. Neither one fixes anything on its own.

Where to start

1. Pick your top revenue products and run the checks above against those pages first. Fixing the long tail before the products that matter is how teams run out of time.
2. Ask ChatGPT, Google AI Mode and Perplexity the ten buying questions your shoppers actually ask. Record whether each product appears as a product card, a brand mention, or not at all. That is your baseline.
3. Fix one gate first: verify the search crawlers can reach the pages, then close the biggest product data gaps on that list.
4. Expose shipping and return terms in structured data and put ship time on the page.
5. Rerun the ten questions after the fixes ship and compare against the baseline. Appearances in answers, AI referred sessions and orders are three different numbers, so report them separately.

Frequently asked questions

Does this only matter for Shopify stores?

No. The assistants read every store the same way, and Shopify's scorer accepts any URL. A custom headless stack faces the same five areas as a Shopify Plus store.

How is this different from traditional SEO?

Traditional SEO ranks whole pages against a query. AI assistants retrieve specific passages and structured fields and match them against the narrower questions they generate from the shopper's request. A page that ranks first in Google can still be left out of an AI answer if its product facts are too thin to match. The two disciplines overlap on technical access and share most of the content work.

What is an llms.txt file and do we need one?

An llms.txt file is a markdown document at yourstore.com/llms.txt that summarizes your site for AI crawlers, lists important product pages and points to your policies. It is a low effort addition and some crawlers read it. Google says it is not required for its AI features, so treat it as optional and fix verified access and product data problems first.

Should we fix product data before investing in AI visibility tracking?

Do the first pass on your top revenue products, then start tracking. Tracking a thin or inaccurate catalog tells you that you are missing from answers without telling you what to change. We cover that decision in should you fix product data before spending on AI visibility.

The five areas are not new. They are how AI shopping answers have been choosing products all along, and they are the same product information work that every other channel depends on. If your team does not have time to work through them, a Paz expert finds the gaps across your product pages and feeds, you approve the fixes, and the platform publishes them and tracks whether your products start showing up. Talk to us.