Scot Wingo has extensive knowledge in e-commerce disruption. Twelve years ago, he capitalized on an earlier online shopping boom, leading to an IPO for his software company, ChannelAdvisor.
Currently, he’s pursuing a new e-commerce venture tailored for the AI era after dedicating a significant portion of the past decade to running an on-demand car service. His new venture, ReFiBuy (short for research, find, buy), aims to create software that assists consumer brands and retailers in navigating an online landscape where AI shopping agents make purchases as often as human consumers.
Wingo recently discussed how generative AI and AI agents are transforming online shopping and the potential implications for e-commerce giants and traditional retailers. Unsurprisingly, as someone developing a major new business in this field, Wingo remains optimistic about how quickly and intensely new AI shopping tools might disrupt established e-commerce players and legacy retailers. The conversation has been edited for brevity and clarity.
Interviewer: What prompted you to re-enter the e-commerce space?
Scott Wingo: Companies dealing with large language models (LLMs) like ChatGPT are capturing consumer interest, which translates to consumer distribution. Once you achieve that distribution, the question becomes how to monetize it. Notably, when Perplexity launched its first shopping feature, it was clear that this would be the next big wave in e-commerce, and I couldn’t afford to miss it. My team and I have a unique perspective as subject matter experts, understanding both the successes and pitfalls of e-commerce to date.
What challenges are you referring to?
E-commerce growth in the U.S. has significantly slowed. Depending on the data source, e-commerce accounts for about 15% to 20% of total retail sales, a figure that has remained quite stable.
However, comparative figures from other regions suggest this number should be higher. Europe and Asia show much higher e-commerce penetration. The stagnation of user experience in the U.S. could be contributing to this disparity. While I’m also an Amazon fan, it seems they have halted significant retail innovation and are primarily focused on profitability now.
Given your mention of Amazon, how do you think it will cope in this evolving landscape?
Amazon has its retail business, a third-party seller marketplace, and advertising. However, the ad layer adds little value to the customer experience. Imagine an AI service that operates over Amazon’s backend, effectively eliminating the need for the Amazon interface. If consumers start using an AI chat or agent instead, that could siphon off a substantial portion of Amazon’s $60 billion in profit margins. That’s the first major issue.
The second challenge is being relegated to the backend, which would mean Amazon has to pay a fee to the AI company to remain relevant. They’re traditionally the ones collecting fees, so this shift presents a conflict regarding who gets to collect those fees.
Additionally, if I were in the shoes of ChatGPT or [OpenAI CEO] Sam Altman, I would consider offering a lower take rate for product listings compared to Amazon’s fees. A seller might prefer listing with us if it came at a lesser cost, potentially destabilizing Amazon’s marketplace revenue.
This approach could erode Amazon’s advertising and marketplace revenues, which are critical profit centers for them.
When you discuss building on Amazon’s infrastructure, are you referring to AI search engines or shopping agents that provide a new customer entry point for online shopping? How will shopping via AI agents differ from our current experiences?
Initially, the experience will be multimedia, potentially incorporating voice, typing, or other methods. Voice interactions are likely to become prominent due to advancements in that technology.
Your AI assistant will manage various tasks and have proactive conversations about your needs. It could remind you of upcoming events and suggest household replenishments based on your preferences. Essentially, it’s designed to be a comprehensive personal assistant, streamlining the purchasing process.
Who might be involved in the backend operations of this system?
The assistant will decide between options like Amazon or Walmart based on parameters you set for value or convenience. For instance, you might specify a budget for purchasing necessities, allowing for efficient processing of orders.
In terms of new purchases, the assistant will ask insightful questions to understand your preferences and make tailored suggestions. For a significant purchase, like a pickleball paddle, it would factor in your urgency and preferred payment methods.
What role can Amazon’s new AI functionalities, such as its Rufus AI shopping assistant, play in its competitive strategy?
While they’re a step in the right direction, I believe tools like ChatGPT will outperform them. The challenge for Rufus is that for it to replace existing search experiences, it would risk sacrificing a significant portion of Amazon’s advertising revenue, which they rely heavily on.
Based on your discussions, do you sense that Amazon or Walmart recognizes the extent of the threat from these developments?
I haven’t spoken with Amazon directly, but Walmart has publicly stated they are open to working with shopping agents. This approach may provide them an opportunity to compete more effectively with Amazon, as they’re less affected by shifting the transaction front end.
What key advice do you have for retailers or brands regarding necessary operational changes?
My candid recommendation is to embrace these agentic systems, as they will become paramount. However, many data pools currently in retail are unprepared for this shift.
The industry lacks standardization for product catalogs, leading to inconsistencies across retailers. An agent-based system requires more detailed and coherent data. Typically, product searches use fewer words than AI queries, which are more complex and nuanced. We aim to assist retailers in assessing their data readiness and identifying areas of strength and improvement, ultimately providing support where necessary.
In your opinion, how rapidly will these changes unfold?
I started conceptualizing this idea in Q4 2024, but in the past month, the urgency has escalated significantly for major retailers. Internal pressures are mounting, especially as the landscape of inbound traffic shifts dramatically. The e-commerce sphere has seen increasing tension, prompting companies to respond swiftly.
