Sidd Motwani, Ian Anderson, and Shivaditya Sinha spent years building the behavioral intelligence infrastructure that powers Spotify’s recommendation engine. The system, called Vector AI, is designed to predict a person’s intentions and next actions, rather than relying solely on a person’s past actions. This powers approximately 90% of Spotify’s recommendations to its 800 million users.
Now, the three are implementing a similar system for e-commerce at their new startup Malachyte. The company announced Thursday that it has raised $10 million in seed funding to expand distribution and hire more product and commercial leaders.
Malachyte was founded on the belief that most online stores treat shoppers the same. This means that personalization is primarily determined by purchase history, demographic segmentation, or logged-in customer profile. This means first-time visitors often see the same general storefront as everyone else, while existing shoppers receive recommendations primarily based on what they’ve previously purchased, rather than what they currently need.
The startup wants to change that by building a real-time intent-aware shopping experience. Its platform uses what it calls “double-headed vector AI” to predict what products shoppers will want next, learn their general preferences, and continually fine-tune them based on their real-time behavior.
“[Our]system starts forming before the first click, using the context available the moment the page loads. Within a single session, it actually reads both the preferences and what someone is trying to accomplish right now,” Motwani, CEO of Malachyte, told TechCrunch.
“With no account or history required, a search for ‘heavy-duty boots’ followed by two clicks on steel-toed boots is enough to move work pants and gloves up the page and push dress shoes down. With each additional action, your profile becomes sharper, so the longer a user stays, the more relevant their experience will be on their next visit.”
Motwani argues that retailers already have the most valuable source of customer intelligence, but rarely leverage it in real time.
“Hovers, clicks, scrolls, narrowing searches, and adding to cart are all signals that most systems either don’t react to in the moment or aggregate into segments overnight. We read them continuously, so each action makes the user vector more confident about both their preferences and current intent,” he added.
He also believes that contextual signals remain underutilized.
“A visitor who calls in from an email link at 11 p.m. and the same visitor who uses a laptop in the morning are in different mental states, and most systems treat them the same way,” Motwani says.
The company has been developing and testing the technology since 2024, working with more than 20 enterprise customers in the travel, grocery, and retail industries before finally focusing on e-commerce.
That platform went live in the fall of 2025 with Fun.com. It has been generally available to Shopify merchants through native integration since June 2026, but larger retailers can integrate the technology through an API.
Looking ahead, Motwani says the bigger opportunity lies in integrating merchandising and marketing on the same understanding of customer behavior.
The funding round was co-led by Bessemer Venture Partners and Gradient, with participation from Harpoon Ventures.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
