The Retailers That Win AI-Mediated Shopping Will Do Three Things

This upcoming holiday shopping season is already shaping up to be unpredictable. Inflation, geopolitics and fickle consumer sentiment are conspiring against retailers’ best-laid plans. But here’s one prediction that requires no crystal ball: this will be known as the year artificial intelligence truly broke through into online shopping.

Study after study has shown that around two-thirds of U.S. shoppers use tools like ChatGPT, Gemini or Perplexity to manage purchases in the past three months. Among Gen Z the figure is up 80 percent.

Retailers need to compete fiercely for online visibility in a market where discovery, comparison and conversion can now happen in a single chat window. Page 1 of Google’s results is no longer the most valuable piece of real estate in e-commerce. Today, it’s the five or six product recommendations that a chat agent offers up to a consumer. A retailer that isn't optimized to be featured on that list will miss the moment.

Based on Koddi’s research into the changing shape of e-commerce, here are three factors that the winners will get right as agentic AI becomes an ever-bigger part of the shopping journey.

1. Understanding what shoppers do and don't want from AI.

It’s easy to get swept away by the AI hype. However, as lululemon’s former chief information officer recently cautioned, the gap between promise and reality is still substantial for many retailers trying to deploy AI across their operations.

That’s why retailers must understand where their customers are on their own journey with AI. Koddi’s study, for instance, shows that consumers treat AI like a personal shopper who can scour the internet for products they might want and offer them up for approval. Around 75 percent of U.S. consumers told us they’re comfortable with AI helping them choose what to buy, and 64 percent like the idea of AI suggesting brands they wouldn’t normally consider. However, shoppers are far more reluctant to hand over their credit card to an agent, with only 32 percent having let an AI service complete a purchase or booking on their behalf. That suggests that retailers should focus investments in the near term on visibility and integrations within chat services rather than overbuilding their digital capabilities for an era of fully agentic commerce that may be slower to arrive than they imagine.

2. Connecting data, decisions and execution.

Retailers’ first-party data is among their most valuable assets, but agentic AI is accelerating the speed at which consumers move. The successful retailers will be those whose data infrastructure can keep up. We’re entering an era in which the window to get in front of the right consumer at the right moment is narrowing. Retailers need robust systems to connect their first-party consumer data with information on pricing and inventory, then package it in ways that are understandable to AI agents.

In addition, retailers need to be open to new ways to compete and collaborate with other players in the industry. Our research showed that 81 percent of commerce media networks and 73 percent of advertisers said an independent platform that pulls data and inventory from multiple retailers into a single, AI-ready feed would be very valuable. Put simply, data that AI agents can read and access is data that drives sales.

3. Proving results from advertising.

Commerce media networks have proven themselves to be highly successful, with most major retailers investing heavily in developing their platforms. As AI reshapes consumer paths to purchase, it will be vital for retailers to continue investing in measurement solutions that can demonstrate real return on investment to advertisers. Proving incrementality is the key. Gopuff, the instant delivery company, showed how it could be done on its commerce media network using randomized test and control groups to quantify true sales lift from advertising exposure. Brands were able to use this information to optimize their campaigns and saw a more than 40 percent increase in incremental purchases on average.

For retail media networks, a continued focus on measurement will be critical as advertisers seek to understand performance in agent-mediated commerce environments. Studies, including our own, show an overwhelming majority of commerce media networks are already responding and prioritizing investment in this area.

There's no silver bullet that will prepare retailers for agentic AI. But there is a clear theme across all of the above: retailers need to focus on ensuring their data and technical infrastructure can meet the real opportunities AI is creating today, while being judicious in preparing for a fully agentic future.

The market is already moving, and retailers need to build quickly to stay ahead of the competition.

Eric Brackmann is vice president, commerce media at Koddi, an industry-leading commerce media technology provider.