Woman sitting in chair interacting with transparent screen showing AI recommended home decor items

Wayfair CTO maps agentic path across digital and brick-and-mortar commerce

IT Management Comments Off on Wayfair CTO maps agentic path across digital and brick-and-mortar commerce

Wayfairโ€™s signal is not that agents will replace the shopping journey; it is that leadership is trying to define where autonomy helps and where human judgement must remain in control. For CIOs and transformation leaders, that boundary is the real management decision. If AI is used mainly for discovery, research and fit-checking, the operating goal shifts from โ€œfully automated commerceโ€ to โ€œbetter conversion, fewer returns and higher customer confidence.โ€

That has funding and governance consequences. Tooling for customer-facing agents, supplier data enrichment and employee productivity cannot be treated as separate experiments if they all depend on the same content, identity and policy controls. The next question is who owns the rules for what an agent can say, recommend or infer, especially when product data is incomplete, supplier-provided or inconsistent across channels. In a retailer with both digital and physical touchpoints, weak decision rights will quickly become a customer-experience issue.

The most interesting operating-model move is the idea that AI is embedded across domains rather than isolated in a central โ€œAI team.โ€ That can accelerate adoption, but only if there is enough central discipline around standards, model oversight, legal review and exception handling. Otherwise, local enthusiasm can produce fragmented use cases, duplicated spend and uneven risk posture. Leaders should ask how reusable capabilities are governed, how model outputs are audited and how frontline teams are trained to challenge AI when it is wrong.

For portfolios, the trade-off is clear: broad enablement creates momentum, but the value case must be measured in operational terms, not novelty. The practical next questions are whether AI is reducing catalog friction, improving assisted selling and shrinking cycle time for internal work; which metrics will prove that; and how quickly the organisation can stop low-value experiments without losing trust in the programme.

Wayfair isn’t handing over full control of shopping decisions to AI, but agents are changing how the retailer’s shoppers discover and evaluate products, according to CTO Fiona Tan. As with other enterprise sectors, retailers continue to inject AI across their operations. Home furnishings retailer Wayfair appears to have an early start, based on Tan’s comments at last week’s Momentum AI conference, hosted in New York City by Reuters. In a sit-down discussion with Tan, Reuters reporter Arriana McLymore noted that Wayfair’s CEO Niraj Shah previously described AI as a significant growth driver for the company that would boost efficiency. In its most recent financial results, for the first quarter ended March 31, Wayfair reported a net loss of $105 million on $2.9 billion in revenue. That compares with a net loss of $113 million on $2.7 billion in revenue for the year-ago period. The reduction in net loss included a decline in operating losses, though that was not attributed to the use of AI.   Tan added that the technology is not entirely new to her company, which has long used predictive machine learning and AI. She said Wayfair has a history of applying data to its operations and, more recently, has begun working with generative AI for in-house and external needs. “We’re a digitally native company โ€ฆ you see us investing in the customer experience, as well as internal operations,” Tan said.

Management on board with cultivating AI

Wayfair adopted a pragmatic investment approach to AI while applying the tech broadly with support from senior management, she said. This includes offering AI tools to employees that let them spend more time connecting with suppliers rather than pulling data, Tan said. Expected uses for AI continue to evolve across industries, including what retailers predicted even one year ago. “At the time, the thinking was that we were going to move toward a very autonomous shopping experience,” Tan said. Rather than have agents handle most of a customer’s shopping decisions, Wayfair is putting AI to use as a boon to product discovery and research on customer-facing platforms, she said. “Our internal services are callable to AI agents.” Wayfair’s goal is to increase customer engagement on its platforms with this approach. For example, AI agents can assist customers with remodeling or redecorating projects. Earlier this year, Tan spoke at the NRF Retail Big Show in New York City, pointing out how AI could warn customers when a purchase such as a sofa might not fit where intended. AI agents can also learn from customers’ decisions not to complete purchases.

E-commerce roots offer an easy shift to AI

When asked how Wayfair’s use of AI compares with its retail peers, Tan remained diplomatic but affirmed its spread. “I think everybody’s leaning into it,” she said. Tan pointed out that Wayfair, founded as e-commerce company CSN Stores in 2002, has digitally native roots that gave it the data and content infrastructure to support AI-enabled offerings. For example, generated images let customers visualize products in real spaces. “If you were to do that before, it would require a lot of 3D rendering, cost and time that just wasn’t practical,” Tan said. Given the scale of Wayfair’s operations, AI may have a hand in other time-saving efforts. Wayfair works with some 20,000 suppliers who offer more than 30 million products, she said. “Each supplier is different, and so how they manage and what they send is also quite different.” Faced with such a volume of goods, AI helps Wayfair update its catalog, letting suppliers add products very quickly without requiring as much information as needed previously, Tan said. At the same time, those updates also ensure customers receive accurate, robust details on products.

Synchronizing AI with the real world

After Wayfair opened its first physical location in 2019, its digital resources also fed its real-world stores. “Going from a digitally native retailer to having brick and mortar, one of the advantages is that all of our systems โ€ฆ whatever utilities that I made available online, it’s available in the physical retail store,” Tan said. That includes letting customers engage smoothly with the company in digital and real-world formats. “It should be a very seamless move between assets,” she said. Wayfair continues to encourage its staff to further embrace the use of AI tools, with something akin to a leaderboard and employees discussing what they have done recently with AI, Tan said. This includes the domain team, the legal team and the accounting team, all having access to AI tools to encourage new ideas, she said. “I don’t think of AI as separate, as in having an AI team. I have an applied science team, but for the most part, just from our history, we have AI embedded within every domain that we have across Wayfair,” Tan said.
https://www.informationweek.com/machine-learning-ai/wayfair-cto-maps-agentic-path-across-digital-and-brick-and-mortar-commerce

Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.