AI Shopping Creates an Opportunity for Retailers to Reimagine the New Storefront

Something strange is happening in e-commerce. Consumer spending is up. Online time is up. Yet traffic to major retail websites is declining. The answer is actually quite simple: shoppers are changing how they research and where they buy.

Their starting point, which used to be about discovery, is now where they're making decisions. Many may never reach the retailer's website at all if purchasing can be done in a single space.

At PayPal, we process global e-commerce volume across more than 400 million active accounts. What we see in that transaction graph is telling: purchase journeys are completed further from the retailer's owned properties. Consumers are increasingly browsing, comparing, and shortlisting inside artificial intelligence tools, often without visiting a product page. By the time they do visit a retailer’s property, the decision is already made.

The retailer's front door has moved. Most retailers haven't noticed yet.

The Website is No Longer a Place Consumers Visit. It's a Signal of AI Consults.

For decades, digital retail ran on a predictable model: invest in traffic acquisition, optimize the on-site experience, and close the sale. The owned-and-operated storefront was the center of gravity. Brands spent billions perfecting their homepage and their customer’s journey because that's where the conversion happened.

That logic holds only if shoppers begin where you want them to. Increasingly, they don't.

AI systems are collapsing discovery, comparison, and recommendation into a single interaction. They surface winning products based on price, reviews, availability, and transaction history signals,  not brand affinity or homepage design. The retailer’s website becomes one possible place a consumer might start among many. The brands with the strongest signals in the AI's environment win, not necessarily those with the best direct-to-consumer (DTC) experience.

A beautiful storefront that nobody arrives at is not a competitive advantage.

Who's Exposed — and Who Isn't

DTC brands built entirely around direct homepage traffic are the most exposed. Their entire model assumes ownership of the first touchpoint: own the relationship, own the data, own the conversion. When AI intermediates that first touchpoint, that plan falls apart immediately.

Retailers with cross-merchant transaction data are quietly positioned on the other side of this shift. The advantage of seeing what a shopper buys from them, what they considered, what they rejected, or what they bought elsewhere reveals category intent, competitive switching behavior, and true purchase propensity. That's the data AI shopping agents will rely on to make optimal decisions on behalf of consumers. Whoever has the data will become the new commerce infrastructure, not just a channel where commerce takes place.

The uncomfortable truth is that most retailers are still measuring the wrong things. Metrics like homepage visits, time-on-site, and add-to-cart rates were built for a world where retailers owned the journey. The new metric that needs to be measured is share of AI-referred purchase completions, and yet almost nobody is tracking it.

What to Do About it

Three moves separate the retailers that will define the next storefront from those that will be disintermediated by it.

The first is to stop treating your website as the only conversion surface. The storefront is now any touchpoint at which a consumer is ready to purchase. That can be social commerce, shoppable media, marketplace integrations, or AI-native surfaces. Brands that insist on their website as the primary path to conversion are ceding surface area to competitors who show up earlier in the journey, at the moment of intent rather than after it.

Second, connect media directly to the transaction. The traditional model places brand advertising at the top of the funnel and hopes the consumer makes their way down. AI-driven shopping compresses that funnel, sometimes into a single exchange. Media must now function as a potential point of purchase, not a consideration to trigger. That means measuring closed-loop conversion against actual transaction data as well as reach and frequency.

Finally, build cross-merchant signal partnerships before the window closes. This is the move that's hardest to reverse-engineer later. First-party data alone is a partial view. Retailers that forge the right data partnerships now with payments platforms, ad networks, and peer ecosystems will have a durable moat that will become difficult for others to overcome.

The Honest Answer

Right now, AI owns the front door.

That's a sentence most retail executives aren't ready to say out loud. But the transaction data is already showing it. The brands that act on it earliest will get something rare: the chance to define what the storefront means in the next era rather than spend the following decade trying to adapt to a definition someone else wrote.

Traffic to your homepage declining is a signal. The question is whether your organization reads it as a threat or as an instruction.

Dr. Mark Grether is the senior vice president and general manager of PayPal Ads, a smarter advertising platform built across PayPal, Venmo, PayPal Honey, PayPal Open, publishers, and merchants everywhere.