AI commerce is here. Meet Joa on Cop.

Shopping often starts with a question. A hoodie for colder days. A gift for someone difficult to buy for. A pair of sneakers that fits a particular look. Joa, Cop’s shopping assistant, makes that question the starting point of the store.

We built the web experience around a continuous conversation: describe what you want, explore products, refine the selection, and move toward checkout. The assistant does the work of navigating the catalog and preparing the purchase. The customer stays in control of the decision.

From a prompt to a real product

A useful shopping assistant needs a connection to the store itself. Joa’s product tools use Cop’s catalog to look up offers, available variants and stock. A suggestion can become a selected size and a cart item instead of ending as a paragraph of generic advice.

This changes the shape of discovery. Customers can explain the occasion or ask a follow-up question without translating every preference into another filter. The product information still has to come from the merchant, and the final price comes from checkout.

Checkout belongs in the conversation

For a signed-in customer on the web, Joa can prepare an order review inside the chat. Items, delivery details and the total are brought together before the customer approves the purchase. Payment entry can appear within the thread; a bank may still require a separate verification step.

The conversation is connected to the checkout system. A clear purchase confirmation authorizes the next step. If the cart or total changes, the customer needs an updated review and must confirm again. An order is only described as placed after the checkout system reports success.

Autonomy with a clear decision point

“Autonomous commerce” is often used to describe an agent that handles the entire shopping task. Our approach gives the assistant useful work to do: find products, check options, prepare the cart and guide checkout. Spending still requires the customer’s clear approval.

That distinction makes the experience easier to trust. The assistant can move the task forward without turning an ambiguous reply into permission to buy. Delivery changes, payment issues and updated totals remain visible parts of the conversation.

One journey, with the store behind it

The interface is only the visible part. Product data, cart state, delivery information, payment handling and order confirmation all have to agree. We designed Joa’s web checkout around those connections so that the conversation can lead to a real transaction.

The next step for commerce is a store that understands what a customer is trying to do and helps them complete it. Joa brings that idea to Cop, one shopping conversation at a time.

Try Joa on Cop or explore the Cop project.