AI Shopping Traffic Fails Mass Retailers But Delivers Elite Spenders

Key Shopping Keyword: ai shopping traffic

Artificial intelligence is delivering retail’s best customers, but no one on an earnings call is calling it a blockbuster yet. The second quarter brought the first clear performance data on AI-driven commerce, revealing a powerful paradox. The technology is acting as a near-perfect filter for high-value buyers, but it has failed to become a meaningful source of new traffic for even digital-first retailers, according to PYMNTS.

Earnings Calls Reveal AI's Retail Paradox: High Intent, Low Volume

Etsy CEO Kruti Patel Goyal set the tone last week. She told analysts that traffic from AI agent platforms "remained under 1% of Etsy’s total," a negligible sliver of overall site visits. DoorDash CEO Tony Xu echoed the point on his company's call, stating that "agentic order volume from AI partners remains low across the board."

The counterpoint is what makes this data significant. That tiny stream of AI-originated traffic is not just converting, it's outperforming. For Etsy, it converts at a higher rate with a larger average order size. According to Adobe data cited in the report, revenue per visit from AI referrals ran 37% higher than non-AI traffic as of March 2026. These shoppers spend 48% more time on-site and browse 13% more pages. Shopify's data aligns: AI-referred traffic tripled year-over-year in Q2, and orders through those channels tripled right along with it.

This creates the central tension. Retailers are seeing a channel with elite customer quality that, for now, doesn't move the volume needle. The industry's scramble to integrate with platforms like Google's Gemini is a bet on a future funnel, not a celebration of a present-day windfall.

Cracking the Code on AI's Super-Shopper Conversion Rates

The reason for these elevated metrics is straightforward. AI shopping traffic represents pre-qualified, high-intent demand. A user asking a chatbot for "a handcrafted leather journal for a law school graduate" has moved past generic browsing. The AI parses that specific request and surfaces relevant results, effectively doing the discovery legwork that a user would normally perform through multiple searches and site visits.

This makes AI agents the ultimate pre-screening filter. They deliver users who are already deep in the consideration phase, if not ready to buy. As we've seen in other sectors like AI deal shopping, the technology excels at parsing complex needs and comparing options, but the final human approval remains critical.

The trade-off, however, is significant. This hyper-efficient, query-driven model may come at the expense of the serendipitous discovery and impulse buying that fuel a large portion of retail revenue. AI is excellent at answering a question, but it is not yet designed to inspire a question the shopper didn't know they had.

The Stakeholder Split: Why Retailers, Shoppers, and Platforms See Different Math

Each player in this ecosystem views the same 1% traffic figure through a completely different lens.

• Retailer CFOs see an intriguing but nascent channel. The high conversion rate and order value are attractive from a customer acquisition cost (CAC) perspective. However, until the volume scales, it remains a promising experiment rather than a pillar of growth strategy. Their primary goal, as noted in additional context from The Economic Times, is to keep the final transaction on their own site to retain customer data and loyalty. Ulta Beauty's Josh Friedman explicitly stated the company prefers customers complete purchases on Ulta's website, not within an AI platform.

• Shoppers are separating discovery from trust. PYMNTS Intelligence data shows a clear divide: 48% of online shoppers used AI to research their most recent purchase, but only 49% would let an agent complete that purchase. For non-regular AI users, that trust level plummets to 3%. The convenience of AI stops at the checkout page.

• AI Platforms (Google, OpenAI, etc.) view any successful referral as a critical proof point. They need to demonstrate their agents can drive real commerce. Even low-volume, high-converting traffic validates the utility of their tools and lays groundwork for future, more integrated monetization. OpenAI has already pivoted, ending an "Instant Checkout" feature to focus on discovery and partner integrations.

This dynamic explains why, as reported in Etsy's own case, users who find products through ChatGPT typically return to Etsy's own site to buy. The handoff is broken.

What the 1% Traffic Threshold Means for Main Street and Marketplaces

For a platform like Etsy, where inventory is unique and purchases are often considered, this high-value, low-volume traffic is intrinsically valuable. It brings exactly the kind of deliberate, appreciative buyer the marketplace wants. As covered in our analysis of Etsy's strategy, their AI focus is on enriching buyer profiles and personalizing discovery, which this traffic directly rewards.

For mass-market retailers, however, sub-1% traffic is a rounding error. It forces a strategic choice: do they optimize deeply for this high-intent niche, or invest in broadening AI's appeal to drive larger volumes? The current data suggests a near-term future where AI agents serve as a premium, high-conversion channel for specific verticals and complex purchases, not a universal traffic source for all of retail.

The Multibillion-Dollar Friction Point After the AI Handoff

The blockage isn't in discovery. It's in everything that happens next. As DoorDash's Tony Xu pinpointed, every AI shopping assistant eventually must hand off to a person or system to check real-time inventory, route an order, and manage fulfillment. That "last-mile" of the transaction, final price comparisons, applying loyalty points, selecting shipping options, and the tactile act of payment, remains riddled with friction that AI cannot yet navigate autonomously.

"Every AI shopping assistant eventually must hand off to a person to check what’s actually in stock, route an order to the right merchant and load a bag correctly. That handoff remains unsolved industry-wide.", Tony Xu, DoorDash CEO

This is not solely an AI problem. It's a legacy retail systems problem. For an AI agent to truly "close the sale," it needs deep, real-time API access to inventory, pricing, and logistics data across a fragmented landscape of retailers. Today's infrastructure isn't built for that. The retailer's own website often remains the only place where all those threads can be tied together, which is why they fight to pull the user back.

When AI Shopping Traffic Will Actually Move the Needle

Significant growth in AI shopping traffic volume hinges on two developments.

First, AI agents must evolve from "search and send" tools into true purchasing agents capable of cross-retailer arbitration. This means not just finding a product, but guaranteeing the best final price including taxes and shipping, applying available coupons, and managing the checkout across disparate systems. This level of agency requires a trust leap consumers have not yet made.

Second, the breakthrough use case may not be a single item purchase. It will likely be complex, multi-item purchases (like weekly groceries) or managing recurring household needs. An agent that can reliably replenish pantry staples across multiple merchants and optimize for cost and delivery would create indispensable utility, transforming traffic from a trickle to a stream.

For now, the earnings call transcripts provide a reality check. AI is not revolutionizing retail traffic patterns in 2026. Instead, it is carving out a high-stakes, high-value niche. The retailers winning are those, like Etsy and Ulta, who are integrating their carts and loyalty programs into these AI streams, ready to capture the high-intent user the moment the AI's job ends. The watch item is no longer whether AI can find a product, but which platform or retailer solves the industry's last great handoff problem.

The Bottom Line

• Retailers are investing in AI integrations despite minimal current volume because the customers it brings are exceptionally high-value spenders.
• The 37% higher revenue per visit from AI referrals shows the technology's potential to transform profitability even at small scale.
• This paradox reveals a strategic shift where quality of traffic is becoming as important as quantity for e-commerce growth.