Box just dropped a major play in the enterprise AI race, unveiling Box Automate - an operating system for AI agents that could reshape how companies handle unstructured data. CEO Aaron Levie announced the platform at Boxworks 2025, positioning the cloud storage giant as a safer alternative to foundation model companies rushing to capture enterprise workflows.
Box CEO Aaron Levie has been watching the AI wars unfold with a particular kind of confidence. While foundation model companies race to build bigger, more powerful systems, Levie's betting on something different - context and control.
At Thursday's Boxworks conference, Box unveiled Box Automate, what the company describes as an operating system for AI agents in enterprise environments. It's the latest salvo in Box's aggressive AI push that started with their AI Studio last year, followed by data-extraction agents in February and search capabilities in May.
The timing isn't coincidental. Just this week, Anthropic rolled out direct file uploads to Claude.ai, creeping closer to Box's enterprise territory. But Levie sees an opening where others see competition.
"We're in the era of context within AI," Levie told TechCrunch in an exclusive interview. "What AI models and agents need is context, and the context that they need to work off is sitting inside your unstructured data."
Box Automate tackles what Levie calls the automation gap - the vast majority of enterprise workflows that touch unstructured data. While companies have automated structured database operations through CRM and ERP systems for years, documents, contracts, and marketing assets have remained stubbornly manual.
"Think about any kind of legal review process, any kind of marketing asset management process, any kind of M&A deal review," Levie explained. "We've never been able to bring much automation to those workflows because computers just haven't been good enough at reading a document."
The platform's key innovation lies in workflow segmentation. Rather than deploying one massive AI agent to handle entire business processes, Box Automate breaks tasks into discrete chunks with specific handoff points. A submission agent might handle initial document intake, then pass work to a separate review agent, each operating within defined parameters.
This approach addresses what Levie sees as a fundamental limitation in current AI systems. "We've already seen some of the limitations even in the most advanced fully agentic systems like Claude Code," he noted. "At some point in the task, the model runs out of context-window room to continue making good decisions."
The segmented approach also tackles enterprise customers' biggest AI fear - agents going rogue. "You don't want to have an agent make some compounding mistake where, after they do the first couple 100 submissions, they start to kind of run wild," Levie said. Box Automate lets companies decide exactly how much autonomy to grant each agent before requiring human oversight or handoffs.
But Box's real competitive advantage might be what it's built over the past two decades - enterprise-grade security and permissions. "When an agent answers a question, you know deterministically that it can't draw on any data that that person shouldn't have access to," Levie emphasized. "That is just something fundamentally built into our system."
This security foundation becomes crucial as foundation model companies eye enterprise markets. While tools like Claude.ai now accept file uploads, they lack the granular access controls and data governance that enterprise customers demand at scale.
"If you think about what enterprises need when they deploy AI at scale, they need security, permissions and control," Levie said. "They want their choice of AI models, because one day, one AI model powers some use case for them that is better than another, but then that might change, and they don't want to be locked into one particular platform."
Box positions itself as the Switzerland of enterprise AI - providing the infrastructure layer while connecting to "every leading AI model that's out there." It's a calculated bet that enterprises will choose flexibility and security over the allure of cutting-edge foundation models.
The announcement comes as enterprise AI adoption accelerates. Companies are moving beyond simple chatbot implementations toward more complex workflow automation, creating opportunities for platforms that can bridge current AI capabilities with enterprise requirements.
Box's bet on the 'era of context' represents a fascinating counter-narrative to the foundation model hype cycle. While competitors chase increasingly powerful AI capabilities, Levie is building the enterprise plumbing that makes AI actually usable at scale. Whether this approach wins out depends on how quickly enterprises prioritize control and security over raw AI horsepower - but with 20 years of enterprise trust already banked, Box might just have the patience to let the market come to them.
