Institutions coordinate constantly. Information moves between teams, responsibilities pass from one unit to another, specialists combine different forms of expertise and decisions travel through layers of authority before becoming action. Much of this coordination can appear remarkably smooth when work is familiar. People know whom to contact, responsibilities are relatively stable, communication channels are available and established routines tell participants what should happen next. Under these conditions, an institution may have good reason to believe that coordination is one of its strengths.

But routine coordination can conceal how much of that success depends on the routine remaining intact.

A coordination stress test asks a different question. Instead of observing whether people and units coordinate successfully under ordinary conditions, it examines whether coordination remains available when some of the assumptions that normally support it are weakened. What happens when a problem suddenly crosses organisational boundaries? When two legitimate instructions conflict? When authority shifts unexpectedly? When several units face pressure simultaneously and the communication capacity on which they normally depend becomes constrained?

The objective is not to create chaos. It is to discover whether apparent coordination reflects a resilient institutional capability or a favourable operating environment.

Consider a public organisation in which two departments regularly work together. Under normal circumstances, their interaction may be highly effective. Experienced employees know one another, responsibilities are understood and recurring issues follow familiar pathways. Now imagine that an unusual problem emerges that belongs partly to both departments but clearly to neither. The established sequence no longer provides an obvious answer. Each unit possesses relevant knowledge, yet neither knows who should integrate it or who has authority to resolve disagreement.

The institution still contains the necessary expertise. What has been stressed is the architecture that allows expertise to become coordinated action.

This distinction matters because having the right components is not the same as being able to coordinate them under strain. An institution may possess excellent teams, strong information systems and competent leaders while still becoming cognitively fragmented when ordinary coordination assumptions fail. Routine performance may never reveal this weakness because familiar work repeatedly follows pathways that participants already understand.

One useful stress condition is therefore cross-boundary disruption. Many institutional problems fit established jurisdictions reasonably well. Others do not. They cut across policy domains, levels of government, professional communities or organisational units. Under normal conditions, experienced staff may compensate for these boundaries through informal relationships. A stress test can ask what happens when those familiar bridges are unavailable or when the problem crosses boundaries that existing relationships do not cover.

Does information still reach the people who need it? Can responsibility be established without excessive delay? Can different forms of expertise be integrated rather than merely exchanged? If coordination depends almost entirely on a small number of boundary-spanning individuals, pressure may expose a capability that is far less institutionalised than everyday performance suggests.

A second stress condition is conflicting instruction. Coordination is easiest when authority points in one direction. Real institutions sometimes face situations in which legitimate objectives compete, different leaders issue incompatible priorities or formal responsibilities overlap. The diagnostic question is not whether disagreement exists. It is whether the institution possesses ways of remaining coordinated while disagreement is being resolved.

A fragile system may respond by fragmenting: each unit follows its own instruction, information stops travelling and the institution begins producing several locally rational but collectively incompatible responses. A more resilient coordination capability can preserve enough shared awareness for actors to recognise the conflict, make it visible and prevent separate parts of the organisation from behaving as though their own instruction were the only one that existed.

A third test concerns authority redistribution. Institutions often coordinate partly because people know who can decide. If authority is temporarily transferred, decentralised or made ambiguous, familiar coordination pathways may stop working. The stress test asks whether coordination can adapt when the usual authority map changes. Do people know how to discover where responsibility has moved? Can information follow the new decision structure? Can units reorganise their interaction quickly enough to preserve institutional coherence?

This is particularly revealing because organisations sometimes confuse stable hierarchy with strong coordination. Hierarchy can certainly support coordination, but a system that coordinates only while every authority relationship remains exactly as expected may be brittle. Robust coordination requires some capacity to re-establish connection when the familiar map changes.

A fourth stress condition is simultaneous demand. Many coordination arrangements work because difficult cases arrive one at a time. A small group of experienced people can resolve exceptions, translate between units and repair misunderstandings. Their work makes the wider architecture appear effective. When several cross-cutting problems occur simultaneously, however, those compensatory mechanisms may become overloaded.

This can expose hidden dependence on scarce coordinators, limited communication bandwidth or sequential processes that cannot cope with parallel demand. Again, the lesson is not simply that performance deteriorated. Almost any institution will slow under sufficient pressure. The important question is where coordination begins to lose coherence and why.

A useful coordination stress test therefore needs a capability-survival criterion. Perfect coordination under strain is unrealistic. Some delay, disagreement and local adaptation may be entirely appropriate. The relevant question is whether the institution retains enough coordination to maintain a shared understanding of the problem, move critical information across boundaries, establish workable responsibility and prevent different parts of the organisation from unknowingly undermining one another.

This makes the failure signal more informative than a single performance score.

Perhaps information reaches every unit but too late to affect decisions. Perhaps everyone possesses the same information but interprets responsibility differently. Perhaps authority is clear but expertise cannot reach the authority holder. Perhaps units coordinate internally while becoming disconnected from one another. Perhaps an informal network compensates successfully until simultaneous demand overwhelms it. These are different coordination failures, and each reveals a different architectural vulnerability.

The diagnostic value lies precisely in this specificity. “Coordination failed” is much less useful than knowing what stopped coordinating with what, under which pressure and through which broken connection.

Digital platforms and AI can strengthen coordination by accelerating information flows, identifying dependencies and helping actors maintain shared situational awareness. But they can also introduce new assumptions into the coordination architecture. If several units rely on one shared platform, what happens when that platform becomes unavailable? If an AI system helps route cases or recommend responsibilities, what happens when its recommendation conflicts with human judgement or formal authority? Technology can increase routine coordination while simultaneously creating new points whose failure becomes visible only under strain.

A coordination stress test should therefore remain diagnostic rather than experimental. Its purpose is to place meaningful pressure on existing coordination capability and observe whether that capability remains institutionally usable. Deliberately comparing alternative authority structures, communication architectures or coordination configurations in order to establish which arrangement performs better would answer a different question. Testing whether coordination survives pressure is not the same as experimentally determining which coordination architecture is causally superior.

The distinction protects the usefulness of the test. We do not need to redesign an institution every time we want to learn something about its fragility. Sometimes it is enough to ask what happens when ordinary supports are no longer guaranteed and to observe which dependencies suddenly become decisive.

An institution has robust coordination not merely when its parts work together under familiar conditions, but when they can preserve enough connection, shared awareness and workable authority to continue acting coherently when those conditions become difficult. A coordination stress test makes the difference visible. It reveals whether everyday cooperation reflects an institutional capability capable of surviving strain, or whether the institution has been coordinating successfully because routine conditions have quietly been doing much of the work.