Live shopping platform Whatnot has acquired AI startup Shaped to power real-time recommendations, targeting the hardest problem in live commerce, helping shoppers find the right product when inventory and demand change by the minute. The lesson for retailers is that recommendations only convert if they keep pace with reality.
Live commerce broke the recommendation engine. Whatnot just bought the fix.
Live shopping platform Whatnot has bought AI startup Shaped, a machine learning company built around real-time recommendations and search. The deal targets the single hardest problem in live commerce. Helping a shopper find the right product when the inventory, the auctions and the demand are all changing by the minute.
Traditional ecommerce runs on a stable catalogue. You know what is in stock, you know the price, and the recommendation engine has time to think. Live shopping is the opposite. Auctions end in minutes or run for hours, stock appears and vanishes, and the algorithm has to react in real time or it recommends something that is already gone. Shaped's technology is built for exactly that.
Shaped founder Tullie Murrell and close to a dozen engineers and researchers are joining Whatnot, where Murrell will lead a new Applied AI Research group. Financial terms were not disclosed.
Whatnot has the scale to justify it. Launched in 2019, the platform recently passed 1 billion orders, raised $225 million in a Series F that valued it above $11 billion and added 20 million buyers in a year.
Why it matters
Live and social commerce is coming to Australia whether local retailers are ready or not, and the discovery problem Whatnot is solving is the same one every retailer faces. The lesson is not that you need to buy an AI company. It is that recommendations only convert if they keep pace with reality. A product suggestion that ignores what is in stock, on promotion or trending right now is a wasted impression.
Whatnot has passed 1 billion orders since launching in 2019, on a platform now valued above $11 billion. Source: TechCrunch.
What to do about it
Check whether your recommendations reflect reality. If they push out-of-stock or full-price items during a sale, they are costing you conversions.
Treat discovery as a conversion lever, not a nice-to-have. Helping the shopper find the right thing fast is where live and social commerce is won.
Watch live shopping seriously. It is small in Australia today and it will not stay that way. Learn the mechanics before you need them.
Feed your systems good product data. Real-time recommendations are only as good as the inventory and pricing signals behind them.
The retailers who make discovery fast and accurate will own the next wave of commerce. The ones still serving stale recommendations will watch shoppers scroll past.