Shopify reports 29.5% GMV growth and discusses the rise of AI-agent shopping

A Shopify representative said the company’s third-quarter GMV was $119 billion, up 29.5% year over year; revenue reached $3.7 billion, up 32%; and free-cash-flow margin was 22%. He said MAC, Clinique, Balmain, and Lululemon moved to Shopify during the quarter, and that the company was beginning to see more commerce through agentic services and personal agents embedded in them.

The representative described agentic shopping as a developing area that remains small relative to Shopify’s overall volume: Shopify’s total GMV last year was $400 billion, he said. Still, traffic from AI agents was up 2.6x year over year and orders were up 3.3x. Conversion using the Shopify catalog was 2x higher than with general AI search, which the representative characterized as effectively scraping. He said early AI beneficiaries include “long-tail” products, including those outside the top 100 categories.

A host said he had made his first agentic purchase: he asked Dot in ChatGPT to find and buy a raincoat, then paid with Shop Pay. The Shopify representative said connecting Shop to personal agents can give them access to purchase history for recommendations; Shop Pay also simplifies payment, and the Shop app can track packages. He characterized Shop Pay as a kind of identity system for agents connected to Shop Pay.

In the representative’s view, agentic commerce is moving from product discovery toward delegated purchases. He said agents may be especially useful when shoppers have not chosen a specific brand or need an item with particular attributes: an agent can compare many products with the shopper’s needs. He cited a specialized Roborock robot vacuum, WeThings smartwatches, and a car seat that fits three children and the back of a sedan as examples.

The representative advised merchants to check Shopify’s knowledge center to see what agents can learn about their products and to provide relevant attributes, such as whether a raincoat works in cold weather. In his view, complete product information can make a product more likely to surface in agent responses. A host added that sizing details and precise measurements—including conversions between metric and imperial units—could help an agent match an item to a user’s measurements. The representative also said merchants selling across borders should clearly state which products can be shipped to each country: restrictions on particular types of fish, for example, may depend on duties and tariffs.