In this episode of , Kevel CEO James Avery sits down with entrepreneur Scot Wingo, to discuss the rise of agentic commerce, AI-driven shopping experiences, and what retail media networks must do to survive the shift.
Retail media networks are facing a fundamental threat. As AI-powered platforms like ChatGPT and Perplexity integrate shopping directly into their interfaces, consumers can now research, find, and buy products without ever visiting a retailer’s website.
In this episode of Unlocking Retail Media, Avery speaks with Wingo — founder of ChannelAdvisor and now CEO of refiBuy — about how agentic commerce is compressing the shopper journey and what that compression means for advertising dollars and retailer-owned media.
Here are the five biggest takeaways from their conversation.
“I’ve been at this technology stuff for 30 years and I’ve been through the mobile revolution, the whole thing. I’ve never seen anything this fast.”
— Scot Wingo, ChannelAdvisor & refiBuy
Wingo traces the rapid evolution of AI-driven shopping from late 2022 to today. Early versions of ChatGPT lacked web access and were prone to frequent hallucinations. By mid-2024, however, web search capabilities made the platform genuinely useful for research. Soon after, product cards appeared, enabling consumers to compare options directly inside the interface.
Now, with checkout fully integrated, the entire research-to-purchase journey can happen without leaving the AI platform. “In July, the Financial Times leaked they’re working on a checkout,” Wingo reflected. “Then that [capability] just came out on 9/30.”
ChaptGPT is on track to reach a billion users faster than any technology in history — four years faster than the next closest competitor. Walmart has already integrated with ChatGPT’s checkout, enabling customers to complete purchases without leaving the platform.
This shift isn’t theoretical. It’s already happening, and the adoption curve is steeper than anything retailers have experienced before.
If a customer checks out on ChatGPT, they never see ads on your retail media network.
“OpenAI can definitely go build an ad platform,” Wingo mused. “Which, we all think they will. But that’s very much more a Google ad platform than it is something for each retailer.”
Retailers have spent years building onsite advertising businesses that depend entirely on site traffic. When AI platforms intercept the shopping journey, they capture both the customer relationship and the advertising opportunity. While OpenAI may build its own ad system, it won’t offer retailers the same control over targeting, creative, or measurement that their owned networks provide.
While some in the retail media space dismiss this shift as limited to basic reordering behavior, Wingo strongly disagrees. He was clear about the risk of that mindset:
“What bothers me is giving people advice to ignore this — because it’s really bad advice.”
According to Wingo, most retail product catalogs have degraded to the point where navigation filters and search boxes are now “hopelessly broken” and “totally useless.”
Amazon’s search experience — once best-in-class — has declined to the point where critical specifications often exist only in product titles. Applying filters frequently eliminates nearly all results, leaving shoppers with “two weird things that are not even in the category you're looking for.”
At the same time, AI platforms are building their own product databases by crawling the web and structuring product information more effectively than traditional e-commerce systems.
Wingo outlined the three layers AI systems use to form product knowledge:
Retailers that fail to clean up their product data, allow AI bots to crawl their sites, and optimize metadata risk disappearing entirely from AI-driven shopping experiences. Wingo noted that refiBuy often uncovers surprisingly basic visibility issues — including IT teams blocking GPT bots by default.
"We find the CEO doesn't know, and they'll look at the person next to them, ‘are we blocking those?’ And, finally the IT guy's like, yeah, ‘I thought they looked bad. I didn't know what they were.’"
Amazon’s ad revenue is approaching $60 billion annually, but Wingo argues that this growth has come at a significant cost to the user experience.
"It started to really negatively impact the user experience to the point where there's so many sponsored products,” he said. “It’s almost like I'm looking for a product and they buried it in ads."
Sam Altman, CEO of OpenAI, has been explicit about avoiding Google-style sponsored listings. Instead, he has pointed to Instagram and TikTok as models for advertising that feels additive — ads that surface relevant products naturally, without overwhelming organic results.
Wingo also highlighted Instacart’s approach, which surfaces promotions between the find and buy stages — such as offering a $20 coupon if a shopper adds another item to their cart. These moments are contextual, helpful, and conversion-driven, without degrading the experience.
Retailers must ask themselves a difficult but necessary question: does our ad strategy improve the customer experience, or make it worse? If the answer is worse, AI platforms will win by default.
Some critics argue that agentic commerce is overhyped because e-commerce represents just 16% of total retail sales. Wingo acknowledges the statistic but emphasizes the unprecedented speed of adoption.
"There's always this asterisk that I've never seen anything grow this fast,” Wingo said. “[ChatGPT] has grown exponentially fast…they’re the fastest thing to a billion users and four years ahead of everybody else.”
Already, roughly half of all e-commerce happens on marketplaces, largely driven by Amazon. AI platforms are well positioned to capture an outsized share of that digital commerce, particularly in categories where speed and convenience outweigh the need for physical inspection.
Even high-consideration purchases benefit from AI assistance. Wingo cited researching a $60,000 watch: while most shoppers won’t complete that purchase inside a chat interface, AI excels at education, comparison, and narrowing choices.
“ChatGPT, interestingly, said ‘these are out of stock everywhere. Do you want me to check on this every day?’” Wingo shared.
For lower-consideration items — such as golf balls, used cars with specific criteria, or collectibles — the path to purchase is already dramatically compressed. Retailers who dismiss this behavior as niche risk repeating the same mistake early e-commerce skeptics made two decades ago.
Retailers can’t stop agentic commerce — but they can adapt. Wingo’s advice is straightforward: stop thinking about retail media networks in isolation and refocus on the customer experience. Fix the product catalog. Ensure AI bots can crawl your site. Structure and maintain accurate metadata. Build loyalty programs that give shoppers a reason to return.
For onsite advertising, make ads additive rather than extractive. Surface relevant products at moments that feel helpful, not disruptive. Study what Rufus and Sparky are doing at Amazon and Walmart — these aren’t just chatbots, but efforts to rebuild the shopping experience from the ground up.
And don’t dismiss agentic commerce as limited to reordering or low-consideration purchases. Wingo was unequivocal:
“If you’re in retail media networks, you should definitely be concerned about this.”
Walmart has already integrated with ChatGPT checkout, and more retailers will follow. The question isn’t whether AI will change shopping — it’s whether your retail media network will still exist when it does.
For more insights like these, tune in to the full episode of Unlocking Retail Media, the podcast where Kevel CEO James Avery sits down with industry leaders and innovators shaping the future of retail advertising.
Listen to this episode with Scot Wingo here.
