Aaron Levie, CEO of Box, stated on July 24, 2025, that the fundamental intelligence for AI agents across every vertical is currently "just sitting in the best AI models right now waiting to be tapped into." His social media post emphasized that the right combination of "domain understanding, tool calling, context engineering, and UX will unlock so many use cases," leading to market expansion as AI models continue to improve.
Levie, a prominent voice in enterprise technology, views 2025 as "day one of agent adoption in the enterprise." He explained that while AI models have existed for years and assistants like ChatGPT emerged more recently, the current evolution points to agents that can "fundamentally go do work for you." This shift is expected to enable a significant increase in demand for previously unaddressable tasks, potentially leading to a "10X or 100X increase" in certain categories of work if the cost of deployment becomes affordable.
Box, under Levie's leadership, has been actively pursuing an "AI-first" strategy, launching a new AI platform featuring advanced agents for search, deep research, and data extraction. These capabilities aim to help businesses gain deeper insights from their content without moving data across platforms, integrating with tools like Microsoft 365 Copilot. Levie believes that companies with the richest, most proprietary datasets are best positioned to succeed in this new era of AI agents.
Despite the immense potential, Levie maintains a pragmatic view, acknowledging current limitations such as "compound errors in chain-of-agent setups, poor search recall, and the risk of surfacing confidential information." He stresses the importance of focusing on clear use cases that can deliver immediate return on investment with high accuracy, urging organizations to adapt their data architectures to leverage AI effectively. The ultimate goal, he noted, is to hide complexity and deliver better outcomes for users.