Malachyte is turning anonymous shopper behavior into a live storefront signal
Malachyte says it can infer intent from a shopper’s live behavior in under 200 milliseconds, without waiting for a login, saved history, or declared preferences. The startup applies the same behavioral logic its founders built at Spotify, where they say the recommendation infrastructure powers more than 90% of recommendations across 800 million users and 1 billion items.
In commerce, that means the storefront can keep rewriting itself during the session. Search results, product listings, banners, and recommendations change as a shopper scrolls, hovers, filters, clicks, or backs out. Merchants can also adjust the model with their own priorities around inventory, margin, promotions, and discovery.
Seller takeaway: AI personalization is becoming an in-session decision layer, so product attributes, inventory state, and promotion rules need to be usable in real time, not just in overnight merchandising logic.
August 19, 2026 462 3 1
