ai recommendations Former Spotify engineers Sidd Motwani, Ian Anderson and Shivaditya Sinha have raised $10 million to apply the recommendation infrastructure they helped build at the music streaming giant to online retail. Their new startup, Malachyte, is aiming to make e-commerce personalization more real-time and intent-aware, rather than relying mainly on past purchases or broad customer segments.
ai recommendations
From Spotify’s recommendation stack to shopping
According to the company, the founders spent years working on the behavioral intelligence systems behind Spotify’s recommendation engine. That system, called Vector AI, is designed to predict a person’s intent and next actions instead of focusing only on historical behavior. Malachyte says the technology helps power about 90% of Spotify’s recommendations for its 800 million users.
Now, the same approach is being repurposed for commerce. Malachyte was formed around the idea that most online stores still treat shoppers in a one-size-fits-all way. The startup argues that personalization is often based on what someone bought previously, on demographic segmentation, or on whether a customer is logged in. In practice, that can leave first-time visitors staring at a generic storefront and returning shoppers seeing suggestions that reflect old behavior more than current needs.
What Malachyte says it does differently
The company’s pitch is built around real-time shopping signals. Malachyte says its platform uses what it calls “two-headed Vector AI” to infer what a shopper wants next, learn the person’s broader taste, and adjust continuously as new actions come in during the session.
CEO Sidd Motwani told TechCrunch that the system begins forming a picture of the visitor before the first click.
“[Our] system starts forming before the first click, using the context available the moment the page loads. Within a single session, we build a real read on both preferences and what someone is trying to accomplish right now,” Motwani said.
He gave an example of how the system is intended to react in the moment: “A search for ‘heavy-duty boot’ 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 no account or history required. Every additional action sharpens the profile, so the experience gets more relevant the longer someone stays, and again on their next visit.”
Reading behavior as it happens
Motwani’s argument is that retailers already collect a rich stream of customer signals, but rarely use that information instantly. He pointed to hovers, clicks, scrolls, search refinements and add-to-cart actions as examples of useful cues that can be missed if they are only processed later.
“Every hover, click, scroll, search refinement and add-to-cart is a signal, and most systems either never act on it in the moment or aggregate it into a segment overnight. We read it continuously, so each action makes the user’s vector more confident about both preference and current intent,” he said.
He also said contextual information is often overlooked. In his view, the same shopper can be in a different mindset depending on the device, the time of day or the channel that brought them to the site.
“A phone visitor at 11 p.m. from an email link is in a different state of mind than the same person on a laptop mid-morning, and most systems treat them identically,” Motwani said.
Early testing and customer rollout
Malachyte has been developing and testing its technology since 2024. During that period, the company worked with more than 20 enterprise customers across travel, grocery and retail before narrowing its focus to e-commerce.
The platform first went live in the fall of 2025 with Fun.com. Since June 2026, it has been generally available to Shopify merchants through a native integration. Larger retailers can also connect to the system through an API.
Why investors backed the bet
The $10 million seed round was co-led by Bessemer Venture Partners and Gradient, with participation from Harpoon Ventures. Malachyte said the funding will be used to scale distribution and hire additional product and commercial leaders.
The funding also signals investor interest in applying recommendation technology beyond media and entertainment. Malachyte’s thesis is that the same kind of behavioral modeling that helps streaming platforms guess what a listener wants next could make shopping sites more adaptive to the person in front of them at that exact moment.
The broader opportunity
Looking ahead, Motwani says Malachyte sees a bigger opening in connecting merchandising and marketing around one shared understanding of customer behavior. That would mean not just recommending products more intelligently, but aligning how stores present inventory and how they target shoppers across channels.
For e-commerce merchants, the pitch is straightforward: if a site can understand intent in real time, it may be able to surface more relevant products sooner, reduce friction for first-time visitors and make repeat visits feel more tailored without requiring a logged-in profile. Whether retailers adopt that model at scale will depend on how well the system performs in live shopping environments and how easily it fits into existing storefront tools.
Explore more: Blog Our Services Contact Us
Source: Original report
Was this helpful?
Last Modified: August 6, 2026 at 6:38 pm
3 views
