The onboarding problem most organisations face is not about training. It is about the fact that the process exists only in the heads of the people who built it.

The Rebuilding Cycle

There is a pattern that appears reliably in growing organisations: a team of six runs smoothly. At twelve, coordination starts to fray. At twenty, new people are disruptive not because they are incapable but because every person who joins exposes a process that was never documented, a decision that was made informally, or a workflow that relied on the institutional knowledge of one particular person who is now overcommitted.

The organisation responds by investing in onboarding. Longer induction periods. More training materials. More time with senior team members who answer the same questions they answered for the previous five new hires. The new hire eventually reaches competence, and the cycle resets for the next person.

The root cause is not addressed because it is rarely named correctly. The problem is not that new hires need more training. The problem is that the process they need to run does not exist outside the minds of the people currently running it.

What Process Fragility Actually Costs

Research from ks-agents.com’s 2025 analysis of knowledge loss and turnover found that tacit knowledge represents approximately 80 percent of an organisation’s total knowledge base. Only 20 percent of what an organisation knows is documented in a form that survives a departure or a transition. The rest exists in the habits, judgements, and undocumented practices of individual employees.

80%	Of an organisation’s total knowledge base is tacit – undocumented and lost when people leave ks-agents.com analysis, 2025

The same research found that 42 percent of institutional knowledge resides solely with individual employees, meaning that when those employees leave, the organisation loses the ability to perform nearly half of what they did. The direct cost of replacing a knowledge worker is estimated at 150 percent of their annual salary. The indirect cost – the undocumented process knowledge that cannot be recovered – is not measured and therefore never appears in the calculation.

The question is not whether your processes will be tested when someone leaves. It is whether your processes can survive the test without requiring the person who built them to explain them first.

The Difference Between a Process That Lives in People and One That Lives in Systems

A process that lives in people requires those people to be available, correctly trained, and continuously present to keep the process running. When one of them is absent, on leave, or has moved to a different role, the process degrades. Knowledge is passed informally, which means it is passed inconsistently. The next generation of process owners inherits a slightly worse version than the one before them.

A process that lives in systems runs consistently regardless of who is executing it. The routing logic is configured. The approval sequence is defined. The document template is standardised. The new hire runs the process correctly on their first day because the process does not depend on their individual understanding of how things are done around here.

The Fix That Scales

The one-time fix for process fragility is documentation combined with automation. Documentation converts tacit knowledge into explicit knowledge: the approval chain is written down, the routing rules are mapped, the document categories are defined. Automation then ensures the documented process runs without requiring a person to remember and execute each step.

1. Map the process as it actually runs
Not as it should run in theory, but as it runs today. Who does what, in which sequence, with which documents, and under which conditions. The informal shortcuts and exceptions should be captured, not smoothed over.

2. Identify the steps that depend on a single person’s presence
Any step that breaks when one person is unavailable is a fragility point. These are the steps that most urgently need to move from people to systems.

3. Automate the routing and approval logic
The decision rules that currently live in someone’s head become configuration in a workflow platform. Documents route automatically. Approvals trigger in sequence. Exceptions surface for human attention rather than requiring humans to monitor for them.

4. Run the process without the person who built it
The test of a documented, automated process is not whether it works when the expert is present. It is whether it works when they are not. New hires should be able to run the process correctly before they understand why it is designed the way it is.

Flowmono’s AI Workflow Builder is where this kind of process lives after it has been mapped and documented: routing rules, approval sequences, document management, and audit trails run inside a single platform that does not require a person to hold them together.

If your organisation rebuilds its processes every time someone new joins, see what documented, automated workflows look like on Flowmono.