By Steve Merrill, Founder of WRKNG Digital | September 30, 2026
An AI agent won't complete a purchase it isn't sure about. Before it buys, it checks for six signals. Miss them and the agent bails to a cleaner store.
Agents are the pickiest shoppers you'll ever have. A human pushes through a clunky checkout. An agent treats every unknown as a reason to quit. Here's what it needs to see.
1. A matchable product variant
The agent needs to map the request, size 10 in black, to one exact purchasable unit. Vague variant names or variants hidden in scripts lose it right here.
Name variants plainly and expose them in the raw HTML.
2. Confirmed availability
It checks that the exact variant is in stock in machine-readable form before it commits. Blank availability reads as unavailable.
The agent won't guess. It moves on. The Agentic Commerce Protocol is built around verified availability.
3. A stable, consistent price
The price at checkout must match the price it saw on the product. A jump at the last step reads as risk and triggers abandonment.
Keep pricing consistent from feed to page to checkout.
4. Clear shipping information
The agent wants to know shipping cost and timing before it finishes. Surprise shipping fees are a top reason it drops out.
State shipping plainly and early, not at the final screen.
5. A findable returns policy
A returns policy the agent can locate and parse signals a store that's safe to complete on. Hidden policies raise doubt.
One click from the product page is the target.
6. A predictable checkout path
The agent wants a clean route from product to completed order with no surprise fields or steps. Consistency is the signal.
Boring and predictable wins agent purchases. Flashy and unpredictable loses them.
How we chose these signals
These six map to the points where agents actually stall on live stores: variant matching, availability, pricing, shipping, returns, and checkout flow. They're ordered by where in the purchase the agent hits each one.
Further reading
- Agentic Commerce Protocol (OpenAI + Stripe)
- Google Merchant Center product data spec
- Schema.org Offer reference
Frequently Asked Questions
What signals do AI agents check before completing a purchase?
A matchable product variant, confirmed availability, a stable and consistent price, clear shipping information, a findable returns policy, and a predictable checkout path. Each corresponds to a point where agents stall on live stores. Missing any one gives the agent a reason to abandon and buy elsewhere.
Why do AI agents abandon Shopify checkouts more than humans?
Humans have already decided to buy, so they push through friction. Agents haven't committed, so any surprise they can't resolve, like a price jump, a hidden variant, or unexpected shipping, becomes a reason to bail. Agents complete only when the path is predictable and the data is verifiable.
How do I make my checkout ready for AI agents?
Name variants plainly, keep pricing consistent from feed through checkout, state shipping cost and timing early, make your returns policy easy to find, and remove surprise fields from the flow. Then test your top products by confirming an agent could match, verify, and complete each one without guessing.
Want to know if AI assistants can actually find and recommend your store? Get a free AI-visibility read on your Shopify store at WRKNG Digital. We show you exactly what ChatGPT, Perplexity, and Google AI see when they look at your products.
