AI Brings Retailers Their Best Shoppers Without Closing the Sale

Etsy, DoorDash and Ulta Beauty this week each told investors that artificial intelligence is drawing more traffic and higher-quality visits to their sites. The market data backs them up. The checkout button remains a bottleneck.

Etsy CEO Kruti Patel Goyal said on a second-quarter earnings call Thursday that traffic from AI agent platforms made up under 1% of Etsy's total. That traffic converts at a higher rate and carries a larger average order size. It still represents a sliver of the business.

DoorDash CEO Tony Xu said Wednesday that agentic order volume from AI partners remains low across the board. Most AI platforms today prioritize enterprise software and coding tools over physical delivery, Xu said.

Ulta Beauty Senior Vice President of Digital and eCommerce Josh Friedman told Reuters that the company is embedding its shopping cart and loyalty program into Google's Gemini. Customers are finding Ulta products there, he said. The integration is early stage.

Independent data confirms the growth. AI-referred traffic to U.S. retail sites grew 393% year over year in the first quarter of 2026, according to Adobe, which tracks direct online transactions across more than 1 trillion visits. Revenue per visit from AI referrals ran 37% higher than non-AI traffic as of March. Shoppers arriving from AI assistants spent 48% more time on the page and browsed 13% more pages per visit than traffic from other channels, Adobe said.

Shopify President Harley Finkelstein reported similar trends on the company's earnings call Wednesday. AI-referred traffic to Shopify stores tripled year over year in the second quarter. Orders through those AI channels tripled as well. New buyer orders from AI channels are coming in at nearly twice the rate of other channels, Finkelstein said.

Etsy's clearest AI wins sit in search and personalization, Patel Goyal said. The company has built richer buyer profiles covering more than 65 million shoppers. Those profiles feed real-time personalization into search results, home feed content and marketing outreach. Engagement creates a compounding effect – the more a buyer interacts, the more relevant each subsequent visit becomes.

DoorDash's own AI tools are showing results even without partner agent platforms. An ordering agent called DoorDash Ask can build a grocery cart in under two minutes, Xu said. The company has also automated catalog-building for new merchants joining the platform. The unsolved problem is the handoff. Every AI shopping assistant must eventually check real-time inventory, route an order to the right merchant and confirm a bag is loaded correctly. No AI agent reliably handles that chain today, Xu told analysts.

Consumers themselves draw a firm line at payment. The PYMNTS Intelligence report "The Agentic Commerce Deep Dive: Payment Infrastructure and the Path to Trust for Merchants" found that 48% of online shoppers used AI to research their most recent purchase. A separate PYMNTS Intelligence report, "From Assistive to Agentic AI: Consumers Wade Into Autonomous Commerce," found that 70% of consumers are interested in using AI agents for travel planning. 71% said the same for health and wellness. Only 49% would let an AI agent complete a purchase. Among consumers who do not already use AI regularly, trust in AI platforms to handle that transaction drops to 3%.

Consumers will let AI find the product. Fewer will let it click the buy button. That is the same boundary Etsy, DoorDash and Ulta described from the merchant side this week.

The gap between search and purchase is not unique to one company. It shows up in the survey data and in the earnings calls from three of the largest online retailers in three different verticals. The infrastructure to complete the last mile of checkout remains a work in progress.