What is Box Content Retrieval?

Box is building Nvidia’s Nemotron embedding technology directly into its product. At its developer conference on Thursday, the company showed off a new feature called Box Content Retrieval, which lets AI agents answer queries more effectively by relying on metadata created by Nemotron.

How much faster are AI agent queries with Nemotron?

Box says users will see faster, more accurate and less expensive queries. Kari Briski, Nvidia’s vice president of generative AI software, told TechCrunch she expects queries to be answered four times faster, with 20% fewer errors and 90% fewer tokens used.

Who can use Box Content Retrieval and when?

The feature will be available to all Enterprise Advance customers, with the rollout beginning early next year.

Box CEO Aaron Levie said the change alters how agents handle documents. “Instead of an agent having to kind of read through each document, we can send back chunks from the documents that are most relevant to the user’s query or whatever the task is that the agent is doing,” he said.

The feature is an implementation of Nvidia’s Nemotron Embed project, which offers open-weight embedding models designed to improve automated retrieval systems. Those models use large language models to pre-process data so agents can use it more effectively. Major AI labs have often overlooked this area, but Nvidia has made it a focus of its open source model work. Its Nemotron release in July topped the RTEB benchmark, which measures retrieval quality.

Box says the system is most useful when paired with active agents. It cites lawyers as an example, who could more easily find relevant clauses and precedents. Because Nemotron’s embeddings map underlying concepts rather than relying on simple keywords, a search tool can surface related legal themes even when different wording or language is used.

The Nemotron models are freely available, but Box’s system leans heavily on its existing metadata and permission controls. That makes the feature difficult to recreate from scratch. Levie summed up the scope of the work. “We’re probably packing in five different technologies into one simple API,” he told TechCrunch.

Este artigo foi escrito com a assistência de IA.
News Factory APP - notícias agênticas para impulsionar seu SEO e AEO.