How Generative AI Creates Workplace Bottlenecks—and How to Fix Them

Generative AI has fundamentally altered the economics of knowledge work by drastically reducing the cost of generating ideas. Today, any competent professional with access to a chatbot can draft a dozen plausible strategies, memos, product concepts, or marketing plans before lunch. While AI has also lowered execution costs in some cases, the decline hasn’t been as rapid or as significant. Shipping even one of those ideas still demands weeks, months, or years of effort.

This imbalance is already reshaping workplaces: too many initiatives for teams to handle, too many tools to master, and too many priorities to track. Leaders continue to pile on new work because the cost of imagining it has plummeted, but the cost of actually completing it remains unchanged. The result? A new management crisis: in an AI-driven workplace, the bottleneck is no longer ideas—it’s execution.

The Broad Institute’s Dual Challenge—and Solution

The Broad Institute, a leading MIT-Harvard biomedical research center, faced a similar dilemma a decade ago—twice. When the first human genome was sequenced in 2003, the process took over a decade and cost approximately $3 billion. Today, sequencing a human genome can be completed in hours for under $200. This dramatic cost reduction created unprecedented opportunities but also triggered two major crises at the Broad.

Crisis One: Operational Overload

As sequencing became faster, samples moved through the pipeline more quickly than downstream teams could process them. Work accumulated at critical bottlenecks, leading to severe overload. Technicians began losing samples, and the lab’s efficiency plummeted. The solution? A shift from a “push” system—where each stage sends work downstream as fast as possible—to a “pull” system, where each stage only accepts new work when it has the capacity to handle it.

Crisis Two: Idea Overload

Once sequencing itself became cheap and routine, the Broad’s innovation team faced a surge in new ideas. Projects were initiated constantly, but few were ever completed. As described in an MIT case study, the team was "losing the technology leadership position it had worked so hard to gain." The solution mirrored the operational fix: disciplined prioritization of ideas.

The team created a visual project map—using Post-it notes on a wall—to track every active project and its stage in the development funnel. This exercise revealed two critical insights: many projects were redundant, and the team was juggling at least twice as many initiatives as it could realistically complete. To address this, they implemented a "project funnel" on the wall, adding a "hopper"—a holding area where ideas waited until capacity opened up in the funnel.

Within two years, the team reduced active projects by more than half and significantly increased the number of projects that reached completion.

Why Leaders Struggle to Stop Adding Work

The Broad Institute’s solution seems straightforward in hindsight, but it’s rarely implemented in practice because humans are inherently biased toward addition. A 2021 study published in Nature, led by researchers at the University of Virginia and other institutions, found that people overwhelmingly prefer adding new elements to existing ones, even when it leads to inefficiency or overload. This "addition bias" explains why leaders continue to layer on new initiatives despite the clear need to focus on execution.

Key Takeaways for AI-Driven Workplaces

• Execution is the new bottleneck: Generative AI has made idea generation nearly cost-free, but execution remains time-consuming and resource-intensive.
• Prioritization is critical: Visual tools like project funnels and hoppers can help teams identify redundant or unfeasible initiatives before they drain resources.
• Resist the urge to add: Overcoming addition bias requires deliberate discipline—focusing on completing existing projects before initiating new ones.

"In an AI-saturated workplace, the bottleneck is no longer ideas. It’s execution."