AI Fails to Fix "Swivel Chair Work" Problem

For years, "swivel chair work" has described the inefficiency of employees toggling between applications, copying data, and chasing approvals, costing significant time and productivity. Harvard Business Review reported in 2022 that workers switch applications 1,200 times daily, losing nearly 4 hours per week to reorientation, which equates to approximately 9% of total work time. For a knowledge worker earning $100,000 annually, this translates to $9,000 in lost productivity per employee. For a company with 10,000 employees, this fragmentation results in roughly $90 million in annual losses, stemming from the friction of fragmented digital infrastructure rather than poor strategy or underinvestment.

When generative AI emerged, most organizations provided employees with enhanced tools for existing tasks, such as improved search, faster drafting, and sharper summaries. However, these implementations did not address the underlying architecture of siloed systems, approval chains, and status emails. Consequently, AI was integrated into workflows without redesigning them, often leading employees to manage one additional application window. In many cases, AI has merely accelerated the "swivel chair" effect, making the process faster but not fundamentally more efficient.

The initial wave of enterprise AI primarily consisted of generative, retrieval-augmented systems. These systems excel at surfacing relevant documents, summarizing them effectively, and then returning the information to a human for action. While this represents a valuable capability, it still involves a handoff where the AI produces an output that a person must then execute. In a business context, compensation is based on completed work and achieved outcomes, not merely on the generation of outputs. This distinction highlights a critical gap in the current AI implementations.

The next evolution of AI focuses on the completion of work through agentic systems. These systems are designed to autonomously plan sequences of actions, select appropriate tools, execute tasks across different systems, and ensure auditable governance of the outcomes. This capability moves beyond generating outputs to actively completing tasks and achieving business objectives, addressing the core inefficiency of "swivel chair work" by automating the entire process rather than just augmenting individual steps.