AI Tools & Products
6d ago
Box integrates Nvidia's Nemotron for enhanced content retrieval system
Nov 5, 2026
AI Summary
Box has introduced a new feature called Box Content Retrieval, which incorporates Nvidia's Nemotron embedding system to improve query responses. This system aims to deliver faster, more accurate results for users, particularly benefiting enterprise customers starting early next year.
- Box has unveiled a feature named Box Content Retrieval at its developer conference, integrating Nvidia's Nemotron embedding system into its product.
- The new feature is designed to enhance the efficiency of agents in answering queries by utilizing embedded metadata from Nemotron.
- Box claims that users will experience queries that are four times faster, with 20% fewer errors and a reduction of 90% in token usage, according to Nvidia’s VP of generative AI software.
- The system will be available to all Box Enterprise Advance customers, with a rollout planned for early next year.
- Box CEO Aaron Levie highlighted that the technology allows agents to retrieve relevant document segments without needing to read through entire documents.
- The implementation of Nemotron Embed aims to improve automated retrieval systems by using open-weight embedding models, which have been underutilized in the industry.
- The system is expected to be particularly beneficial for active agents, such as lawyers, by enabling them to find pertinent legal information more efficiently.
- While Nemotron models are available for free, the system's effectiveness also depends on Box's existing metadata and permission controls, making replication difficult.
- Levie noted that the system combines multiple technologies into a single API.
content retrievalnvidiaefficiencyagent-firsttokens