Agents Transform Org Charts Into Work Charts

Asha Sharma envisions a future where AI fundamentally transforms organizational structures by enabling an agentic society where the marginal cost of good output approaches zero.

In Sharma's view, we're entering a world where the traditional org chart will evolve into a "work chart" as capable agents handle an increasing number of tasks. "When all of that happens, the org chart starts to become the work chart... you just don't need as many layers," she explains. This shift will prioritize tasks and throughput over hierarchical reporting structures, with both embedded agents (functioning as tools within software) and embodied agents (handling specific roles like lead generation) becoming increasingly prevalent.

For leaders, this means rethinking how work gets organized. Rather than focusing on who reports to whom, the emphasis shifts to how tasks are automatically routed, who's working on what, and how to observe and fine-tune agent performance. Sharma believes humans will always decide how AI is used, but the mechanics of work distribution will fundamentally change.

For individual contributors, this transformation offers the opportunity to expand their skillsets through personal "agent stacks" they bring to work—similar to bringing your own device. When millions of workers become 20% more skilled through AI augmentation, the economic impact could be exponential. The future workplace will likely require less focus on hierarchical communication and more on outward, task-based collaboration.

This shift requires new approaches to planning and strategy. Sharma's team operates in "seasons" defined by secular changes in the industry (the current season being "the rise of agents"), with loose quarterly OKRs and 4-6 week squad goals. Crucially, they leave slack in the system not just for unplanned work but for "the slope"—continuously thinking about how to disrupt their own platform and what investments will make that possible.

Implications for Observability and Evaluation

As organizations deploy more agents, the challenge of reviewing their work at scale becomes critical. Sharma emphasizes that "the same kind of loop" of fine-tuning, self-healing observability, and robust evaluation systems will be essential. Fortunately, systems that manage billions of users already exist, providing models for device management, policies, and access controls that can be adapted for agent oversight.