Beyond the Endless Scroll: Can Intent-Aware AI Solve the Retail Discovery Gap?

For too long, mid-market retailers have relied on a static version of personalization. Most storefronts treat shoppers as members of a demographic segment or rely on historical purchase data, meaning a first-time visitor is often met with a generic experience, and returning customers are chased by products they have already bought. In my view, this is why the 'endless scroll' persists; the discovery problem isn't a lack of inventory, but a lack of predictive AI that actually understands taste and intent in the moment.

As TechCrunch first reported, a new startup called Malachyte is attempting to solve this by applying the behavioral intelligence infrastructure used at Spotify to the world of e-commerce. Founded by Ian Anderson, Shivaditya Sinha, and CEO Sidd Motwani—who previously built the recommendation engine that powers roughly 90% of Spotify's suggestions for its 800 million users—Malachyte has raised $10 million in seed funding. The round was co-led by Gradient and Bessemer Venture Partners, with participation from Harpoon Ventures.

Malachyte's approach centers on "two-headed Vector AI," a system designed to predict a shopper's next action and general taste in real time rather than relying on logged-in profiles or past behavior. Motwani told TechCrunch that the system begins forming a profile before a user even clicks, using context available at page load. As a user interacts with a site, the system continuously refines its understanding. For example, if a shopper searches for "heavy-duty boot" and clicks two steel-toed options, the AI can immediately prioritize work pants and gloves while deprioritizing dress shoes—all without requiring an account or history.

Motwani argues that retailers are sitting on a goldmine of underutilized intelligence. He notes that signals like hovers, scrolls, and search refinements are often ignored in the moment or aggregated into segments overnight. Furthermore, he highlights the importance of contextual signals; a user visiting via an email link on a phone at 11 p.m. is in a different mindset than the same user on a laptop during the morning, yet most systems treat them identically.

Malachyte has been testing this technology since 2024 with over 20 enterprise clients in the grocery, travel, and retail sectors. The platform launched with Fun.com in the fall of 2025 and has been available to Shopify merchants via native integration since June 2026. Larger retailers can access the technology through an API. Motwani believes the ultimate goal is to unify marketing and merchandising through this shared understanding of customer behavior.