At NRF, leaders from PayPal, Home Depot, and Wayfair discussed the rise of agentic commerce and what AI means for the future of shopping. The panel explored trust, merchant control, and how retailers should prepare for AI-driven discovery and transactions.
In a packed room at NRF, Jason Del Rey, founder of The Aisle, led a wide-ranging discussion with Mike Edmonds of PayPal, Angie Brown of The Home Depot, and Fiona Tan of Wayfair. The conversation explored how AI-driven agents are reshaping discovery, decision-making, and transactions across retail.
While agentic commerce has quickly become one of the most talked-about ideas in retail technology, the panel made clear that its impact will be evolutionary rather than instant.
Mike Edmonds, who leads agentic commerce strategy at PayPal, began by grounding the conversation in a shared definition.
Agentic commerce refers to AI-powered agents that assist consumers and businesses throughout the shopping journey. Today, that shows up in two primary ways.
On-surface experiences enhance branded websites and apps, making them more conversational, personalized, and efficient. Off-surface experiences allow product catalogs to be discovered and shopped through AI answer engines and shopping assistants, with humans still making the final purchase decision.
While fully autonomous agent-to-agent commerce is often discussed, Edmonds emphasized that this future is still some distance away. For now, humans remain firmly in the loop.
Angie Brown, Chief Information Officer at The Home Depot, described the current moment as one of experimentation rather than waiting.
Home Depot is investing in both on-site tools, such as Magic Apron and project-based shopping assistants, and off-site discovery experiences where customers increasingly begin their journeys. The goal is to learn alongside customers and understand how shopping behavior shifts from product search to project-based intent.
Fiona Tan, Chief Technology Officer at Wayfair, echoed this approach. For Wayfair, the home category is emotional, visual, and difficult to describe with traditional filters and keywords. AI enables conversational and multimodal discovery, allowing customers to explore styles, generate images using real products, and better visualize outcomes.
Both leaders agreed that retailers must meet customers wherever they choose to shop.
Despite growing interest in AI-native shopping environments, trust remains a critical barrier to widespread adoption.
Tan emphasized that regardless of where a transaction occurs, Wayfair must remain the merchant of record. Fulfillment, customer service, and returns are core to the brand promise, especially for high-consideration purchases like furniture.
Brown reinforced that perspective, noting that shopping does not end at checkout. Service, support, and post-purchase experiences are inseparable from brand trust.
From PayPal’s perspective, Edmonds explained that payments, identity, fraud protection, and dispute management are foundational to making agentic commerce viable at scale. Trust infrastructure will determine how quickly consumers move from AI-powered discovery to AI-enabled transactions.
When asked how retailers should prioritize investments, all panelists returned to fundamentals.
Clean, accurate product data, pricing, and catalogs are essential. Without them, agentic commerce cannot function effectively.
Tan noted that Wayfair’s experience managing millions of products has shown how foundational data quality is to discovery, both on and off platform. Advances in AI now make it easier for retailers to improve catalog quality than in the past, but the work is unavoidable.
Brown agreed, calling data maturity a prerequisite for participating in the next era of commerce.
The panel pushed back on the idea that shopping will soon become entirely autonomous.
Edmonds described shopping as a deeply human experience. While AI can remove friction and automate tedious steps, full autonomy will likely remain limited to specific categories and use cases.
Brown highlighted Home Depot’s brand-specific agents as an example of practical agentic commerce. These tools help customers define projects, understand requirements, and build carts, all while keeping the experience grounded in the brand’s expertise.
Tan added that in high-consideration categories, success is not always measured by conversion. Sometimes the best outcome is helping a customer realize that a product is not right for them, reducing costly returns and dissatisfaction.
In a closing lightning round, the panel addressed what may be getting too much and too little attention.
The idea of fully autonomous shopping is often overemphasized. Removing friction, improving know-how, and enhancing customer confidence are more immediate and valuable goals.
What deserves more focus is the business model. As multiple platforms, agents, and intermediaries participate in agentic commerce, retailers must understand how value is created and monetized without compromising customer experience.
Agentic commerce is not a single destination. It is the next evolution of search, discovery, and shopping.
Retailers that succeed will not bet on one platform or protocol. They will invest in data foundations, protect their role as merchant of record, and experiment across on-site and off-site experiences.
Most importantly, they will treat AI as a tool to serve customers better, not to replace the human elements that make shopping meaningful.
The future of agentic commerce will belong to brands that combine trust, learning, and disciplined innovation.
SOURCE: Retail.news