Retailers Can’t Afford to Become the Warehouse Behind Someone Else’s AI Storefront
Last holiday season gave us the first real evidence of what happens when consumers shop with artificial intelligence assistants. Adobe measured AI-driven holiday traffic to U.S. retail sites growing almost 700 percent year-over-year in 2025, with those visitors converting 31 percent more than other traffic. This season tests whether that compounds. Someone like me, with no idea what to buy a 16-year-old into K-beauty for Christmas, can ask an AI assistant for ideas, narrow them by interests and price, and arrive on a retailer’s product page knowing almost exactly what they want.
For retailers, that is high-intent traffic. It is still a small piece of overall demand, but these shoppers arrive further along in the shopping process because some of the discovery happened before they reached the store. That's why they convert so well.
Therefore, the question becomes what happens if more and more shoppers increasingly rely on AI agents.
A retailer creates value through much more than stocking a product in a warehouse. What you choose to sell and how you merchandise it shapes why someone picks your store and comes back. Retailers have spent years learning those lessons.
Put the same assortment of products into someone else’s AI interface and much of that influence disappears. If I ask an AI assistant for a brown couch under $1,500, I might get a flat list of options from several retailers. The retailer has little control over why one couch appears before another, and competition quickly comes down to price, availability and whatever the AI decides matters most.
That is where retailers need to be careful that they don't become the warehouse behind somebody else’s storefront.
The stakes go beyond margin on the sale. The direct relationship gives a retailer its first-party data and the high-margin media business built on it. Lose the front door and the advertising revenue goes too because advertisers pay to reach shoppers, and the shoppers are standing in someone else’s store.
Retailers should participate in AI shopping wherever consumers find it useful. The harder question is how much of the customer relationship to hand over. People need a reason to shop directly with a retailer, and that increasingly depends on whether its discovery experience can answer the questions they now ask AI.
Participating also means playing offense. AI assistants recommend the retailers they can read and trust, and most retail sites are still not built to be understood by machines. Retailers that make their catalogs machine readable are likely to win an outsized share of the recommendations, much as early search winners did.
And most retail search still struggles with shopper questions, handling a query for an exact television model well but falling apart when I ask for a good television for a bright room under $800, or a gift for someone who likes K-beauty products but owns all the obvious ones. Consumers now expect useful answers.
The next step is to bring that intelligence into the store itself, with systems that understand intent rather than keywords, trained on real shopper behavior. Search, recommendations and conversational discovery will blend, enabling shoppers to move from broad idea to specific product without leaving the retailer’s environment. The same machinery makes retail media feel like discovery instead of clutter, so retailers that invest in one get better at the other.
Retailers should watch AI-referred traffic closely this season. The encouraging part is that the work overlaps: the catalog and intent understanding that make a retailer legible to AI agents also make its own store better at discovery. If AI assistants become a new front door to commerce, retailers need to make sure they're still building a store on the other side of it.
Pat Copeland is general manager of Moloco Commerce Media (MCM), the only AI-native ads engine that enables retailers and marketplaces to activate high-performing retail media networks.
Related story: Why Your Future Top Customer May Never Actually Visit Your Homepage
- Categories:
- Artificial Intelligence (AI)
Pat Copeland, General Manager, Moloco Commerce Media
Pat Copeland is the General Manager of Moloco Commerce Media, where he leads strategic initiatives, product innovation, and customer engagement. By leveraging machine learning, analytics, and scalable cloud infrastructure, his team supports retail partners—ranging from major e-commerce platforms to emerging digital marketplaces—in optimizing their advertising strategies, enhancing customer experiences, and maximizing their monetization potential.
Pat brings 30 years of leadership across top tech companies. At Zendesk, he led global teams building AI-powered customer support tools. At Amazon, he launched and scaled Sponsored Brands into a multi-billion-dollar ad platform. He spent a decade at Google, where he played key roles in Advertising, Research, and Cloud Systems, helping earn the company IEEE’s 2013 Company of the Year and oversaw award-winning products like Google WiFi. Pat began his career at Microsoft, working on operating systems, web services, SQL Server, and Bing.
Pat holds a Master of Science in Computer Science with a specialization in Machine Learning from the University of Southern California and a Bachelor of Science in Computer Science from the University of Arizona.





