Box taps Nvidia’s Nemotron for new agent-first content retrieval system
techcrunch ·
Box has announced a new feature called Box Content Retrieval, which integrates Nvidia’s Nemotron embedding system directly into its core product. This update is designed to significantly enhance the ability of AI agents to answer user queries by leveraging embedded metadata rather than scanning entire documents. According to Nvidia executives, this integration is expected to make queries four times faster, reduce errors by 20%, and cut token usage by 90%. The system utilizes open-weight models that map underlying concepts, allowing for more accurate retrieval even when different wording or languages are used. Box CEO Aaron Levie highlighted that this approach dramatically boosts search relevance for agents by returning only the most pertinent document chunks. The feature will be available to all Enterprise Advance customers starting early next year. While the Nemotron models are open-source, Box emphasizes that its unique combination of pre-existing metadata and permission controls makes the system difficult to replicate independently.
techcrunch