What is the coordination tax?
The coordination tax is the time and cost a business spends passing work between people and systems: chasing, checking, re-keying and waiting. It grows with headcount and volume, and it's often the largest hidden cost in an operation. It isn't caused by hard work. It's caused by manual coordination and by people checking outputs by hand.
Where it comes from
Every recurring workflow has three layers:
| Layer | What happens | Who usually does it |
|---|---|---|
| 01 Coordination | Information moves between people and systems. Someone prepares, routes, checks and commits work. | People |
| 02 Decision | A rule or a judgment is applied. | People or software |
| 03 Verification | The result is confirmed, reconciled or escalated. | People |
In most businesses, layers 1 and 3 take most of the operational headcount. As volume grows, headcount grows with it. That's the coordination tax.
How to spot it
- Work waits in inboxes and queues more than it's actually worked on.
- The same data is typed into more than one system.
- Errors are found in month-end reconciliation, not when they happen.
- Each new customer or market needs another coordinator.
How it's reduced
Not by adding people to the coordination layer, but by rebuilding it. Agents prepare and route work, deterministic checks verify each step before it commits, and people handle decisions and exceptions. Teams that have done this grow capacity without growing headcount the same way. They don't replace their people; they change what their people do.
Why it matters economically
Ronald Coase's Theory of the Firm explains why companies hire more people as they grow: coordinating work internally gets more expensive as volume rises. Our AMEA and AMEE papers ask what changes when that coordination is done by verified agent orchestration.
Frequently asked
Is the coordination tax the same as overhead?
Not quite. Overhead includes rent and tools. The coordination tax is specifically the people-time spent moving and checking work.
Which workflows carry the most?
High-volume, rule-bound ones such as invoices, claims, collections, onboarding and reconciliation. See the use-case library.
Can software alone remove it?
Not if the software still needs a person to check every output. The verification layer has to be automated too, with people kept for exceptions.
How do I measure it in my business?
The free diagnostic gives a first estimate for one workflow, and the Audit measures it properly.
See what this looks like in one of your workflows.
Related: AI workflow automation · AI-native operating model
More explainers: The verification premium · Agent orchestration vs RPA · Verifying AI agent decisions · Revenue per employee