Tacit workflows keep organisations moving. Until they become a risk.
Tacit workflows are often signs of competence. The risk appears when invisible fixes become part of the operating model without being understood, governed, or designed for.
Tacit workflows are not inherently bad. In many cases, they exist because experienced staff have quietly fixed what the formal system does badly. They know who to call, what shortcut works, where the real bottleneck sits, and how to keep things moving when the process on paper is too slow, too rigid, or too detached from reality.
That is often competence, not non-compliance.
When competence turns into invisible infrastructure
The problem is what happens when those workarounds become invisible infrastructure.
What works informally for a few experienced people can become fragile very quickly. Knowledge sits in people’s heads. New starters struggle to learn it. Training becomes inconsistent. Key individuals become bottlenecks. Teams start saying things like, “Jane normally sorts that,” without noticing that the workflow now partly depends on Jane existing.
That is not just a resilience issue. It is an operating model issue.
Why this matters in regulated environments
In regulated environments, it matters even more. If part of the real workflow sits outside formal systems, governance can be weaker than it looks. An issue resolved informally might feel efficient in the moment, but it can also bypass audit trails, blur accountability, and make it harder to see whether the right controls were followed.
Informal resolution is not automatically wrong. The risk is that it becomes normal without becoming visible.
Why redesign and AI efforts often miss the real workflow
It also makes improvement harder. When organisations try to redesign services, introduce new systems, or apply AI to operational processes, they usually start with the official workflow. But the official workflow is often only half true. The real version includes the shortcuts, exceptions, local fixes, and unwritten judgments that staff use every day.
Ignore that tacit layer, and projects start solving the wrong problem.
That is one reason so many change efforts disappoint. The process looked clear enough on paper. The reality underneath it was not.
What this means for machine-assistable work
For organisations thinking seriously about AI, automation, or workflow redesign, this matters more than it first appears. Before you can make work more machine-assistable, you need to understand how it actually works now, including the parts nobody wrote down. Otherwise, you are not modernising the workflow. You are building on top of invisible dependency.
The official process is the map. The tacit workflow is the path people actually walk.
If you only look at the map, you misunderstand the terrain.