Box CEO Aaron Levie discusses how AI agents will drive a decade of enterprise infrastructure modernization, shifting the focus from managing static content to providing actionable, structured business context.

Key Points

• Software engineering demand is expanding across the economy as AI agents make previously unviable projects economically feasible, a phenomenon Levie describes as Jevons paradox.
• Successful AI integration requires reengineering workflows to provide agents with precise, structured data rather than just connecting disparate systems.
• Enterprises must navigate the "trillion-dollar question" of balancing deterministic code for repeatable processes with probabilistic models for adaptive tasks.
• Startups have a competitive advantage in automating unstructured professional services, such as legal, accounting, and audit processes, where no software incumbents currently exist.
• Box is pivoting its strategy to treat AI agents as first-class users by transforming stored enterprise content into accessible, real-time context for automated workflows.

Why it Matters

The transition to an agent-driven enterprise requires a fundamental overhaul of how companies organize and access their proprietary data. Businesses that fail to modernize their underlying infrastructure will struggle to provide the surgical context necessary for agents to function effectively, potentially ceding market share to AI-native startups.