Your Product Feed Is Now Your Ad, Your Landing Page and Your Checkout
Your next customer might buy without ever seeing your store. Is your catalog ready for that?
For years, a product catalog was back-office data. You wrote the creative on top of it: the ad copy, the landing page, the checkout flow. That order is flipping. When a shopper asks an AI assistant for “waterproof hiking socks under $25”, the assistant does not read your brand story. It reads your catalog.
This post walks through what changed, what agents actually read, a 6-point feed audit you can run today, and how the same feed can drive your ads and landing pages.
What changed
Three things landed within about a week of each other.
1. Checkout inside Google’s AI surfaces. Shopify released its Fall ’26 Edition on October 28. According to coverage of the release, shoppers can now buy from eligible stores inside Google’s AI Mode and the Gemini app, using a checkout powered by Shopify. The order lands in your Shopify admin like any other. It is rolling out and switched on by default for eligible stores that sell to shoppers in the US, Canada or Australia.
2. Agents that run your store, not just shop in it. The same Edition introduced Shopify Connectors. Per ChannelX’s report, they work with nine agents, including ChatGPT, Claude, Gemini and Copilot, so merchants can manage products, inventory, orders and performance from inside a chat. The flip side: the same agents are also where your customers increasingly start.
3. More ads in AI answers. On November 4, Search Engine Roundtable reported that Google can now show up to four ads in an AI Mode response for a single query. One caveat: Google has not said whether this is a test or a lasting change, and PPC Land notes that four is the same per-group ceiling classic text ads have had for years. Treat it as a direction, not a settled number.
Put together: the product catalog is becoming the single source for your ads, your product pages and your checkout inside AI agents. In that world, feed quality matters more than creative.
What agents actually read
Nobody outside Google and the AI labs publishes the exact ranking signals, so take this as practical experience, not a spec. But the inputs are not a mystery. An agent comparing products has to work from structured data:
- Titles. The single most important field. “Merino Trail Sock, Waterproof, Men’s 43-46, Navy” is something an agent can match to a query. “Step Into Adventure” is not.
- Attributes. Brand, material, size, color, GTIN, category, condition. Missing attributes mean the product cannot be filtered or compared.
- Prices and compare-at prices. Shopify’s Catalog now passes compare-at prices to AI agents, so your markdowns show up without extra setup. That is great when the data is right, and embarrassing when a stale compare-at price makes a permanent “sale” visible to every agent.
- Availability and variants. An agent that recommends a product and then hits “out of stock” at checkout learns to avoid you.
- Policies. Shipping, returns, privacy and terms. Reportedly, a store missing one of the required policies does not get the in-chat checkout.
A 6-point feed audit
Pick your ten best-selling products and check each one against these six points. Takes about 30 minutes.
- Title test. Read the title out loud without seeing the product. Does it say what it is, who it is for and the one or two attributes people search for? Put the product type and key attribute first, brand and marketing language last.
- Attribute completeness. Fill every field that applies: brand, material, color, size, GTIN or MPN, and the right category. Count empty fields per product. Anything under 100% for the fields that apply is a gap.
- Price honesty. Check that price and compare-at price are both real. Remove compare-at prices on items that have not actually been discounted.
- Availability and variants. Variants should map to real, in-stock options with accurate inventory. Delivery times should match what you really ship.
- Policies in place. Terms of service, privacy policy and return and refund policy all filled in under Settings > Policies. Also check your eligibility view in Shopify for the AI channels.
- Answer-ready description. Add two or three plain sentences that answer what an assistant would be asked: who is it for, what is included, how is it cared for, what is it compatible with. Add alt text and one clean main image.
Score each product 0 to 6. Anything below 5 is a product an agent will probably skip in favor of a competitor with cleaner data.
How the same feed drives your ads and landing pages
Once the feed is clean, it stops being an admin chore and becomes the engine. Here is one fictional product to show how a single row fans out:
| Feed field | Becomes… |
|---|---|
| Title: Merino Trail Sock, Waterproof, Men’s 43-46, Navy | Ad headline and the page’s H1 |
| Attributes: merino, waterproof, 43-46 | Ad description lines and a spec table on the page |
| Price 24.90, compare-at 29.90 | Price badge in the ad and on the page |
| Description answers (who, what, care) | FAQ block on the landing page |
| Policies (shipping, returns) | Trust section and checkout reassurance |
On the ads side, Google says ads in AI Overviews and AI Mode use the same auction and bidding models as standard Search ads, and that Shopping campaigns, Performance Max and AI Max for Search are eligible. All of those run on your Merchant Center data. A strong feed gives them more to work with; a thin one leaves the system guessing.
On the landing page side, this is exactly how I approach AI-generated campaigns: take the catalog, generate ad variants and a matching page from the same fields, and keep them in sync. When the title, price or availability changes, the ad and the page change with it. No copy-paste, no mismatch between what the ad promised and what the page shows.
What to measure
You do not need a new dashboard on day one. Start with five numbers:
- Feed health. Disapproved or limited items in Google Merchant Center, and your attribute completeness score from the audit.
- AI-channel orders. Shopify’s Agentic Storefronts dashboard shows orders, sales and conversions from AI channels. Use the Shopify admin as your source of truth: one Shopify agency warns that GA4 tags may not fire inside Google AI Mode checkout, so these sales can be undercounted in GA4.
- Revenue and ROAS per product, not just per campaign. This shows which listings your feed fixes actually helped.
- Share of orders from AI channels over time, plus average order value compared with your normal checkout.
- Returns and support tickets on products sold through agents. If an agent sets the wrong expectation, it shows up here first.
Re-run the audit monthly on your top ten products. Feed quality decays quietly: new variants, changed prices, expired promotions.
Next step
Want me to check 10 of your product listings? Comment “FEED” on my LinkedIn post and I will send you the audit results.
Sources: Shopify Fall ’26 coverage (Fudge), Shopify Editions Fall ’26 rundown (Coalition Technologies), Google AI Mode four ads (Search Engine Roundtable), Google AI Mode fourth ad slot (PPC Land).