What we know

Box has integrated Nvidia’s Nemotron AI model into its new agent-first content retrieval system. The company claims this system can answer queries four times faster than previous methods, reduce errors by 20%, and use 90% fewer tokens in processing queries. These improvements aim to enhance the efficiency and accuracy of content retrieval within Box’s platform.

Why it matters

By adopting Nvidia’s Nemotron, Box is positioning itself to improve how users interact with and retrieve content through AI-driven technology. Faster query responses and fewer errors could significantly boost productivity and user satisfaction. The reduction in token usage may also lower computational costs and improve system scalability.

What is still unknown

No independent tests have verified the accuracy or efficiency claims. Details about which languages the system supports and when the updated content retrieval system will be broadly available have not been disclosed.