Live Shopping is the TV home shopping experience redesigned for the Internet.
The host goes live on a platform, demos a product, takes viewer questions in the comments and offers a limited-time deal. Viewers purchase with one click, without leaving the stream. The format reduces the distance between discovery and transaction to just seconds.
In China, live shopping will generate nearly $900 billion in sales in 2025, nearly the size of the entire U.S. e-commerce market, the consumer intelligence firm said. fussy Estimated on Friday (July 17) press release. In North America, by contrast, 68% of consumers said they had never purchased anything via social media.
The format that generates hundreds of billions of dollars in Chinese e-commerce has yet to gain any real footing in the United States
The gap is not in the content. It’s about discovery. Thousands of live streams are running simultaneously. Inventory changes by the minute. Auctions end in seconds. A viewer who joins late misses out on product that could have converted. Matching the right buyer with the right flow at the right moment requires personalization at a speed that static recommendation systems cannot provide. This is the problem WhatDirect shopping market value With more than $11 billion, it’s trying to solve it.
Whatnot processes over 500,000 hours of live video and millions of user interactions every week. The company took over on Wednesday (July 15). In the form ofMachine learning start Which builds real-time recommendations and search infrastructure, TechCrunch I mentioned Wednesday. Financial terms were not disclosed.
Founder and CEO form Tully Murielwho worked on recommender systems in dead Before being founded in 2021, Shaped will join Whatnot alongside nearly a dozen engineers and researchers to lead a new applied research group in the field of artificial intelligence, the report said.
Live shopping requires AI to match buyers with broadcasts in seconds
The challenge Shaped must face is specific to direct commerce. Unlike traditional e-commerce platforms where product catalogs are relatively stable, Whatnot’s inventory is in a constant state of flux, according to the report. Buyer interests can change during the broadcast.
“By combining Shaped’s technology with Whatnot’s existing systems, we can deliver recommendations faster, more responsive and more personalized,” said Emmanuel Fuentes, vice president of data and AI at Whatnot, according to the report.
Shaped systems combine customer data with artificial intelligence and machine learning models to deliver personalized search and discovery, the report said.
This combination can help Whatnot suggest the most relevant flows to users based on past purchases and real-time behavioral signals.
Whatnot launched more than 35 new product categories in 2025 and more than 45 in the first half of 2026, with sellers exceeding 1 billion orders, according to the report.
Each new class adds complexity to the detection problem. A shopper who accesses a live feed at the wrong moment — after an item has sold or before the item they want appears — gets little value and may not return.
AI personalization can bridge the gap between content and purchase
the PYMNTS INTELLIGENCE a report “A deep dive into proxy commerce: payment infrastructure and the path to merchant trust“, in cooperation with Visa acceptance solutionsfound that 48% of online shoppers used AI to research their most recent purchase.
Meanwhile, traffic pointed to the United States retail Locations grew 393% year over year in Q1 2026.
Consumers’ appetite for shopping using AI is growing. Real-time AI can show the right flow to the right buyer at the right moment and update it as behavior changes during the session. By acquiring Shaped, Whatnot is betting that solving the discovery problem is what moves direct commerce from a niche format to a mainstream channel.
For all of our PYMNTS AI coverage, subscribe to our daily newsletter Amnesty International newsletter.
