Williams-Sonoma Says Olive Users Convert at 3× the Normal Rate

Retail AI assistants are meant to move beyond answering product questions and toward revenue or completed service. Williams-Sonoma says Olive users convert at three times the normal rate, while revenue tied to the shopping assistant has risen 620% since January. The figures are a strong company-reported sales signal, but they do not isolate Olive’s causal effect on conversion.

Sameer Hassan, Williams-Sonoma’s chief technology and digital officer, gave the figures on the company’s second-quarter earnings call. The company also said engagement with Olive increased 700% since January. Separately, it said personalized ecommerce visits generate roughly nine times the revenue of an average visit, compared with roughly two times in the prior year.

Sales association versus service resolution

Olive’s conversion figure measures purchasing among shoppers who engage with the assistant against a normal rate. Pottery Barn’s newer Otto assistant is being measured differently: more than 70% of its early interactions were resolved without a human handoff, according to Hassan. The two results show different outcomes—purchasing and automated service resolution—rather than directly comparable performance measures.

What Otto is designed to do

• Coordinate furnishings across a room, including matching items such as a sofa and rug.
• Account for details including rug sizing and outdoor-material durability.
• Book a design consultation or route a shopper to a designer when needed.

A retailer result within a broader referral pattern

The assistant metrics arrived with second-quarter net revenue of $1.96 billion, up 6.7% year over year. Adobe has reported a related, but distinct, result: shoppers arriving at retail sites through AI referrals convert 54% better than shoppers arriving from other sources. AI referrals bring people to a retailer, while Olive and Otto operate within the retailer’s own shopping experience.