Malachite: Spotify's AI Revolutionizes E-Commerce

A New Era for Online Recommendations

Sidd Motwani, Ian Anderson, and Shivaditya Sinha, three former Spotify employees, played a crucial role in creating the behavioral intelligence infrastructure of the music platform. This system, known as Vector AI, is designed to anticipate user intentions and future actions, rather than relying solely on past behaviors. It is responsible for 90% of the recommendations made to Spotify's 800 million users.

Today, these three entrepreneurs have decided to transpose this technology to the e-commerce sector with their new company, Malachyte. The startup recently announced that it has raised $10 million in a seed funding round. This investment aims to expand the distribution of their technology and recruit product and commerce experts.

Changing the Game in E-commerce

Malachyte starts from the observation that most online stores treat their customers uniformly. Recommendations are often based on past purchases, demographic data, or existing customer profiles. This means that new visitors often see a generic interface, while regular customers receive suggestions primarily based on their past purchases, without considering their current needs.

Malachyte's mission is to transform this approach by creating personalized, real-time shopping experiences. Their platform uses a system called “dual-headed Vector AI” to predict the products that customers might want, while continuously refining recommendations based on users' real-time actions.

Real-Time Personalization

According to Motwani, CEO of Malachyte, the system begins analyzing a user's preferences as soon as the page loads, well before they click on anything. During a single session, the platform develops a deep understanding of the user's preferences and current goals.

For example, a search for "sturdy boots" followed by a few clicks on steel-toe models may be enough to display work pants and gloves on the page, while excluding dress shoes. This happens without the need for an account or purchase history. Each interaction refines the user's profile, making the experience more relevant as it progresses.

Leveraging Real-Time Signals

Motwani emphasizes that retailers already possess valuable sources of customer intelligence, but they often do not leverage them in real time. Every mouse movement, click, scroll, or addition to the cart is a signal that most systems ignore or process with a delay.

Malachyte, on the other hand, continuously interprets these signals, allowing for the ongoing refinement of the user's profile and a better understanding of their current intentions. Motwani also believes that contextual signals, such as the time of day or the type of device used, are largely underutilized.

Proven Technology

Since 2024, Malachyte has been developing and testing its technology with over 20 enterprise clients across various sectors, including travel, grocery, and retail. The platform was first launched in the fall of 2025 with Fun.com. Since June 2026, it has been accessible to Shopify merchants via a native integration, while larger retailers can integrate it through an API.

Looking ahead, Motwani sees a major opportunity in the convergence of merchandising and marketing, relying on a unified understanding of customer behavior.

Strong Financial Support

Malachyte's recent funding round was co-led by Bessemer Venture Partners and Gradient, with participation from Harpoon Ventures. This financial backing is expected to enable the startup to continue innovating and transforming the e-commerce landscape.