Box and NVIDIA launch AI search and workflow tools for enterprises

Expanding the AI partnership

Box is growing its partnership with NVIDIA to provide businesses with more practical ways to use artificial intelligence. The two companies are launching new tools that help AI systems navigate internal company files and automate repetitive tasks. This move comes as organizations look for ways to integrate AI into their daily operations.

The collaboration focuses on making AI more reliable for professional environments. By combining Box's content management with NVIDIA's hardware and models, the companies want to help users find specific information more quickly. This is part of a broader trend where software companies are trying to make AI useful for specific business needs.

As Analytics Insight reported, the partnership is designed to help businesses move beyond basic AI assistants. The goal is to create systems that can actually coordinate tasks and complete parts of existing workflows. This represents a shift toward more functional, agentic AI in the workplace.

Improving search and retrieval

One of the primary new tools is Box Context Retrieval, which functions as a search engine. It is powered by NVIDIA's Nemotron models, which are specific types of AI models used for processing information. The system is designed to help AI agents find relevant data within large collections of business documents.

Box claims that this technology can significantly improve the performance of these searches. Specifically, the company says the tool could improve search accuracy by 20% and make question-answering up to four times faster. It also aims to reduce token usage—the units of data processed by AI models—by 90%.

These improvements are intended to help AI agents deliver more accurate responses to users. By making it easier for the AI to find the right context, the system becomes more reliable for complex queries. The product is expected to be available to Enterprise Advanced customers early next year.

Automating workflows with language

Box also introduced a tool called Box Architect, which changes how employees build automated processes. Instead of having to manually configure every single step of a business workflow, teams can now use everyday language to describe what they want to achieve. The platform then builds the AI-powered workflow based on those descriptions.

This approach aims to lower the barrier for non-technical staff to use automation. By using plain language, employees can create complex systems without needing deep technical knowledge of the underlying software. The tool is intended to help teams coordinate tasks and complete parts of their existing work more efficiently.

Box Architect is scheduled to enter beta testing for Enterprise Advanced customers in the fourth quarter. This allows early adopters to test the capabilities of language-based workflow creation before it reaches a wider audience. It highlights a move toward more intuitive software design for business automation.

Security and open models

Because businesses are giving AI systems access to sensitive information, security is a major priority for this launch. Box is introducing new threat protection features to monitor for suspicious activity. These include the ability to identify bulk downloads, account takeovers, and large-scale file deletions.

The company is also working on securing the execution of AI agents through NVIDIA OpenShell. This ensures that as AI systems begin to perform more tasks autonomously, they do so within a secure framework. These security features are expected to reach Enterprise Advanced and Shield customers early next year.

Additionally, Box plans to bring NVIDIA's open-weight Nemotron models into Box AI. Open-weight models give customers more options for running their own AI applications. This gives organizations more flexibility in how they deploy and manage their AI infrastructure.

The shift to AI agents

This expanded partnership reflects a larger trend in the enterprise software market. Companies are moving away from AI that simply answers questions and toward AI agents that can take action. These agents are designed to retrieve information, coordinate various tasks, and complete parts of existing workflows independently.

For Box, this partnership strengthens its offering by focusing on enterprise data, automation, and security. It allows the company to provide a more comprehensive suite of tools for large organizations that need to manage massive amounts of information. The integration of NVIDIA technology provides the underlying power for these capabilities.

For NVIDIA, the deal provides a way to bring its models and infrastructure into the software used by large organizations. By embedding its technology into Box, NVIDIA ensures its AI models are accessible to a wide range of business users. This collaboration marks a significant step in making enterprise-grade AI more practical and secure.