Box AI Agents Secure GSA Deal to Assist US Government Agencies
Cloud content management firm Box has secured a significant agreement with the U.S. General Services Administration (GSA), marking a new chapter in its evolution and expanding the reach of its artificial intelligence capabilities into federal government systems. This deal, announced Wednesday, positions Box AI – the company’s “intelligent content management platform” – as a tool for government agencies seeking to leverage AI for data insights and workflow automation.
The agreement falls under the GSA’s “OneGov” purchasing strategy, designed to streamline procurement for federal agencies. By offering standardized, discounted terms, the OneGov initiative aims to eliminate the need for individual agencies to negotiate their own contracts, theoretically accelerating the adoption of new technologies. While the specific discount rates were not disclosed, they are set to expire in August 2026.
Box, once primarily known for its cloud storage solutions, has undergone a notable strategic transformation over the past decade. Following intense competition in the cloud storage market, the company pivoted its focus towards helping organizations extract deeper understanding from their data using artificial intelligence. This shift culminated in the recent introduction of a new suite of AI agents within its broader Box AI data analytics platform, which the company is now confident will find a receptive audience within the federal sector.
According to Box, its platform will empower government users by enabling them to generate complex insights across vast repositories of files and folders, efficiently extract metadata at scale, and automate repetitive administrative tasks through easily configurable AI agents. A key aspect of Box’s approach is its collaborative model: rather than developing proprietary AI models, Box AI integrates and leverages technologies from a diverse array of leading AI developers. This includes models from OpenAI, Google, and Anthropic, with additional options available from IBM, Amazon, Meta, and xAI. The company asserts that Box AI will ultimately support agencies across a spectrum of functions, from AI-powered workflows and compliant content management to e-signature capabilities, forms automation, document generation, and enhanced collaboration.
Box’s foray into federal AI solutions is part of a broader trend of U.S. government agencies exploring and adopting artificial intelligence. The Pentagon, for instance, has already inked deals to integrate agentic AI into its planning and operational processes, while the Internal Revenue Service (IRS) has publicly expressed ambitions to utilize AI to augment or even replace certain enforcement roles.
However, the rapid embrace of agentic AI by government entities comes with a significant caveat. Recent research published by Gartner casts a cautionary light on the current performance of these technologies in the private sector. The analysis suggests that more than 40 percent of private sector agentic AI projects are likely to be canceled by 2027, primarily due to their underwhelming performance. The study revealed that even the best-performing AI agents successfully completed only 30 percent of office-related tasks, with most achieving success rates below 10 percent. This stark reality underscores a critical challenge for government agencies considering the deployment of agentic AI, whether through Box’s new offerings or other solutions: the imperative to meticulously plan and rigorously test these systems to ensure they operate effectively and avoid unintended consequences.
[[]]
As Box brings AI agents to Uncle Sam, a critical question looms: will these digital helpers truly transform government, or just complicate it?