Wayfair CTO: AI cut API development time by 85%

Wayfair isn’t just using AI to sell you a green couch. According to CTO Fiona Tan, the company is wielding generative AI as a scalpel to carve out technical inefficiencies and a paintbrush to personalize shopping — and the numbers are stark. In a recent interview, Tan revealed that leveraging AI allowed engineering teams to create new APIs 85% faster, an aggressive acceleration that turns a typically resource-heavy slog into a rapid sprint. That’s not an incremental improvement; it’s a demolition of old timelines.

The tech modernization push extends to conquering the boogeyman of every legacy business: technical debt. Tan explained that generative AI is being deployed to untangle and modernize outdated codebases and databases that are traditionally neglected because they’re too complex to touch. Instead of mounting a massive, multi-year overhaul, Wayfair is using AI to understand the spaghetti code and refactor it. It’s a pragmatic kill shot against the kind of accumulated digital rot that quietly throttles speed at big retailers.

Beyond the backend, Wayfair is tackling a specific ecommerce curse: unbranded goods. It’s easy to search for a specific Sony TV, but a ‘green couch’ is a fuzzy concept. Tan detailed how multimodal AI now integrates text, images, and customer preferences to move past basic keyword searches toward a richer conversational interface. This is critical when your catalog is full of visually distinct but verbally ambiguous items. The legal team is also an unexpected power user here, wielding AI to sift through customer feedback to sniff out liability risks and product safety issues with far greater speed and accuracy.

Tan sees the upskilling race splitting into two tracks. While technologists get advanced tools, the real unlock is training non-technical staff in legal, HR, and operations. Her logic is sharp: these teams often spot practical use cases that engineers would walk right past. It’s a bet that domain expertise plus a dash of AI fluency will generate more disruptive ideas than a centralized tech priesthood ever could. The endgame is a dual shopping experience where AI handles quick, precise purchases for the impatient, while powering immersive, video-driven journeys for the browsers — killing the one-size-fits-all web store for good.