The Coordination Bill

The coordination bill includes at least six kinds of work:

• Discovery: finding the relevant people, records, and dependencies.
• Interpretation: reconciling different terms, assumptions, and accounts.
• Commitment: specifying who owes what outcome, under which conditions.
• Monitoring: tracking progress and identifying when a commitment is at risk.
• Resolution: deciding what to do when evidence, authority, or incentives conflict.
• Verification and repair: confirming outcomes and correcting failures.

These costs interact. Poorly specified work increases monitoring. Fragmented context increases interpretation. Weak authority boundaries increase review and dispute. A handoff that appears inexpensive at one point may create delay or rework downstream.

Agents may reduce some coordination costs by gathering information, tracking state, requesting missing inputs, and preparing routine actions. They may also add costs: integrating systems, defining delegated authority, reviewing outputs, handling exceptions, and correcting errors.

The economic question is not whether an agent can perform a coordination task. It is whether the total arrangement lowers the cost of an accepted outcome while preserving quality and accountability.

A pilot that counts messages sent or tasks completed can therefore mislead. It may show faster routing while shifting verification to another team. It may reduce meeting time but increase exception repair. Evaluation needs a baseline for the full process: labor, elapsed time, waiting, error, rework, oversight, and the value of outcomes accepted by the relevant parties.

Coordination is not only administrative overhead. It helps determine which work an organization performs internally, which it buys from others, and how responsibilities are divided. When coordinating across a boundary becomes costly, a firm may bring work inside or create a dedicated management layer. When information moves more easily, other arrangements may become viable.

This does not mean that lower coordination cost automatically produces smaller firms or flatter structures. Other costs also matter: expertise, bargaining power, quality control, risk, investment, and the need for shared capabilities. Agentic systems may reduce one part of the bill while increasing another.

The direction and scale of change remain empirical questions. They depend on which coordination costs fall, who captures the savings, and whether the cost of oversight and failure grows more slowly.

Before automating coordination, identify where the work occurs and who pays for it. Trace one recurring commitment from request to acceptance. Record how much effort goes into clarification, waiting, handoffs, review, exceptions, and repair. Include work performed by customers, suppliers, and downstream teams.

Then compare the current arrangement with a bounded pilot. Measure accepted outcomes, not activity volume. Separate labor saved from labor shifted. Track failures and recovery. A lower coordination bill is valuable only when the organization still meets its obligations.

The enterprise’s coordination bill may become more visible as software takes on parts of the work. Whether it becomes smaller is a testable proposition, not an assumption.