AI Shopping Traffic Is Exploding. Most Retailers Still Don't Own the Data Behind It
Generative AI traffic to U.S. retail sites grew 393% year over year in the first quarter of 2026, on top of a 693% surge during the 2025 holiday season, and shoppers who arrive that way spend meaningfully more time on site than any other channel.
That part of the story is good news.
The part getting less attention is that a growing share of that same shopper journey is starting, and increasingly ending, entirely inside an AI platform retailers don’t control, which means the data that used to belong to the retailer never touches their own systems at all.
The Growth Is Real, and It’s Not Slowing Down
Adobe Analytics, which tracks more than a trillion visits to U.S. retail sites, has been documenting this shift for over a year.
AI-referred traffic to U.S. retail sites rose 393% year over year in Q1 2026, and by March 2026 that traffic was converting 42% better than non-AI sources, a complete reversal from a year earlier when AI traffic converted 38% worse.
Revenue per visit from AI referrals ran 37% higher than the non-AI baseline. On engagement specifically, shoppers arriving from AI sources spent roughly 48 percent more time on site and viewed 13 percent more pages per visit than shoppers from other channels.
The growth rate has moderated somewhat as the comparison base gets larger, but the underlying trajectory hasn’t reversed. This isn’t a novelty channel anymore. It’s a fast-growing, high-intent, high-engagement source of traffic that most retailers are actively courting.
The Part That Gets Less Attention: Who Owns the Data
Courting that traffic increasingly means partnering directly with the AI platforms themselves, and that partnership comes with a cost most retailers are still working out.
Etsy, Target, and Walmart have all integrated with Google’s Gemini and Microsoft’s Copilot, on top of earlier partnerships with OpenAI’s ChatGPT, letting shoppers discover and in some cases purchase products without ever leaving the AI platform’s own interface.
Wharton marketing professor Kartik Hosanagar put the strategic risk plainly to Retail Dive: “It also shifts the locus of power.
Whoever controls the agents now has the power.” When a customer’s entire research-to-purchase journey happens inside an AI platform’s chat window, the retailer loses visibility into what that customer searched for, which products they considered, how long they looked at each one, and everything else that used to live in the retailer’s own analytics.
Deloitte’s 2026 Retail Industry Global Outlook found that 81% of surveyed retail executives expect generative AI to weaken brand loyalty by 2027, largely for this reason.
Bain & Company’s research adds a useful counterpoint: consumers currently trust a retailer’s own AI agent three times more than a third-party one to actually complete a transaction, which is exactly the trust gap retailers building their own tools are trying to capture before it closes.
Why This Is a Harder Problem Than a Tracking Gap
This goes further than the well-documented issue of AI referral traffic showing up as unlabeled direct traffic in standard analytics. That’s a measurement problem, fixable with better tracking.
The disintermediation risk is structural: if a shopper’s discovery, comparison, and checkout all happen inside ChatGPT or Gemini, there may be no session on the retailer’s own domain to measure at all, no matter how good the retailer’s tracking setup is.
Nikki Baird, vice president of strategy at retail technology provider Aptos, described what’s at stake when Google showcased Gemini’s commerce tools at this year’s NRF conference: if the platform isn’t sharing everything that led to the discovery, the decision, and the purchase, “the retailer loses so much context because it’s outside of their realm.”
That’s not a pixel problem. It’s a question of whether the retailer ever gets access to the relationship in the first place.
How Retailers Are Actually Responding
The response splitting the industry isn’t whether to show up in AI-driven discovery, most retailers agree they have to be there, but how much of the transaction they’re willing to hand over.
- Walmart’s approach is instructive: rather than routing checkout through a platform-native flow, it embedded its own Sparky intelligence into the AI layer while keeping the transaction anchored inside its own ecosystem. The basket, the customer data, and the relationship stay under Walmart’s control even when the interaction starts on someone else’s platform.
- Amazon has taken a more defensive posture, notably absent from ChatGPT’s product partnerships while investing heavily in its own Rufus assistant to keep discovery and purchase inside Amazon’s own properties.
- Sephora is following a similar pattern to Walmart’s, using AI within external platforms for discovery while tying the experience back to its own app and loyalty program.
The pattern across all three: participate in AI-driven discovery, but fight to keep the transaction, and the data it generates, on infrastructure the retailer actually controls.
What to Actually Do With This
Before signing onto the next AI platform integration, it’s worth treating it as a build-versus-lease decision for the customer relationship, not just a distribution opportunity:
- Map which parts of the AI-driven shopping journey currently touch your own domain, where you can measure first-party click and conversion data, versus which parts happen entirely inside a platform you don’t control
- Weigh any new AI commerce integration against how much of the transaction and post-sale relationship it keeps on your own infrastructure versus routes through the platform’s own checkout
- Prioritize first-party data collection on the portions of the journey you do control, since that’s the data asset that compounds in value as more of the funnel migrates onto external platforms
- Revisit these integrations regularly, since the specifics of what data a platform shares back with retailers are still being negotiated and are likely to keep shifting
This connects directly to the gap between where customers discover you through AI and what your attribution data can actually see, and to the broader question of who gets credit when an AI buys something on a shopper’s behalf.
Both are symptoms of the same underlying shift: a growing share of the customer journey is happening somewhere the retailer doesn’t own.
Why First-Party Measurement Matters More, Not Less, Right Now
The instinct in a moment like this might be that first-party attribution matters less, since so much of the journey is moving off-domain anyway. It’s the opposite.
First-party attribution isn’t a future problem for ecommerce brands, it’s the foundation for measuring, with real accuracy, exactly the portion of the customer relationship you still control.
As AI platforms absorb more of discovery and, increasingly, checkout, the click-based data your own store generates becomes the clearest evidence of what a customer actually did on infrastructure you own, rather than a summary a platform decided to share back with you.
Every AI commerce partnership is a negotiation over how much of that data you keep.
Measuring what you do control, accurately and independently, is what makes that negotiation possible to have on informed terms. If you want to see exactly how much of your customer journey is currently measurable on your own domain, book a live AdBeacon demo.
—-
FAQ
How much has AI referral traffic to retail sites grown?
Adobe Analytics reported AI-referred traffic to U.S. retail sites grew 393 percent year over year in Q1 2026, following a 693 percent year-over-year surge during the 2025 holiday season. By March 2026, that traffic was converting 42 percent better than non-AI sources.
Why can’t retailers see who’s buying through AI platforms?
When a customer’s discovery, comparison, and purchase all happen inside an AI platform’s own interface, the retailer often doesn’t receive the underlying behavioral data, what the customer searched for, compared, or considered, that would normally be captured on the retailer’s own site.
What is retail disintermediation in the context of AI shopping?
Disintermediation refers to the risk that AI platforms position themselves between the retailer and the customer, absorbing the discovery and purchase relationship and reducing the retailer to a fulfillment role with limited direct customer data or brand interaction.
How are retailers protecting customer data while still using AI platforms?
Some retailers, including Walmart, are embedding their own AI tools into external platforms while keeping the actual transaction and customer relationship anchored on their own infrastructure, rather than routing checkout entirely through the platform’s native flow.