“Just Ask Thomas”: Why Person-Dependent Knowledge Costs You Time Every Day

Companies rarely have too little knowledge. Far more often, the knowledge is simply not available where it is needed at that moment. The foreman knows why a process was changed years ago. The colleague in purchasing knows which supplier will make an exception. The design engineer still remembers why an assembly was built differently back then.

The company owns this knowledge. But access to it runs through people. That is exactly what person-dependent knowledge means: knowledge that exists inside the company but is only reachable in practice if you ask the right person.

As long as everyone is there and has time, hardly anyone notices. The information does get found, after all. The event shows up in no error statistics and no ERP report. Yet there is a cost every single time: someone has to work out who might know the answer, reach that person, and pull them out of their work. And afterwards the answer usually ends up exactly where it was before: in the expert's head, and now also in the head of the person who asked. Next time, the process starts from scratch.

That searching for information eats working time is well documented. In Microsoft's 2023 Work Trend Index, 62 percent of respondents said they spend too much of their workday searching for information; 68 percent said they lack enough uninterrupted focus time. A much older analysis by the McKinsey Global Institute from 2012 put the figure for so-called interaction workers at 19 percent of the work week spent searching for and gathering information, explicitly including tracking down the colleague who can help.

The exact numbers do not transfer to every plant. Both studies come from an office environment, and the McKinsey figure is more than ten years old. The observation behind them, however, is remarkably stable: skilled employees spend a noticeable share of their time not on their actual work but on getting hold of information.

And searching a folder or SharePoint is only the visible part. This is information search too: “Do you know who did that back then?” “Why do we actually solve it this way?” “Who knows this customer?” The company's search engine, in that case, is its staff.

A single conversation may cost only five minutes. That is why it looks harmless. But it always ties up at least two people.

A deliberately simple calculation: an experienced employee is asked for help ten times a day, five minutes each. That is 50 minutes of his working time. The ten colleagues also spend five minutes each getting the information. That makes 100 minutes a day, for one single person being asked. How often this happens in your plant, you know better than any study.

And even that calculation underestimates the effect. An interruption costs more than the minutes of the conversation. It fragments the working day of the person being asked. Germany's Federal Institute for Occupational Safety and Health (BAuA) notes in its Stress Report Germany 2019 that disruptions and interruptions during task execution generally involve additional effort and lost time, and are associated with reduced mental recovery. Microsoft's 2025 Work Trend Index measures an interruption by a meeting, email, or chat every two minutes during core working hours; 80 percent of respondents worldwide say they lack the time or energy to do their job.

The colleague's question at the machine does not show up in statistics like these. For the person who is asked all day, it feels exactly the same.

What makes this particularly uncomfortable is who these questions land on. Nobody asks the colleague who joined three weeks ago. The people who get asked are the ones with experience: the foreman who knows every machine. The sales colleague who has looked after the same customers for fifteen years. In other words, exactly the people whose time is most valuable.

This produces a paradox. The more someone knows, the more often their actual work is interrupted because others need that knowledge. The expert becomes the company's human search index. And the company gets used to it, because it works.

It works until the day Thomas is not there. Vacation, illness, the other shift. Then the answer suddenly becomes: “Only Thomas knows that.” A small inefficiency has turned into an operational problem.

In many companies, knowledge loss only becomes a topic when a resignation lands on the desk or a key employee's retirement draws closer. That is understandable, but late. When a company discovers on someone's last working day that nobody can replace that person's knowledge, the dependency usually existed for years. Every question that only this person could answer was an early signal.

The obvious reaction is: Thomas should document his knowledge. In practice, that only works to a limited degree. Thomas does not know which of his thousand experiences will one day matter to a colleague. And after fixing a breakdown, he wants to get on with the next job, not think about which wiki category his experience belongs in. We have described in a separate article why company wikis go stale for exactly this reason: the effort lands on the expert immediately, the benefit arrives later and usually for someone else.

So the better question is not: How do we get Thomas to write more down? It is: How do we make sure a question that has been answered once is no longer a new question the second time?

The sentence “just ask Thomas” carries valuable information. It shows that relevant knowledge exists, that it is present in the company, and that it is not available independently of one specific person. When the same questions regularly land on the same people, a knowledge map draws itself. What matters is not the theoretical question “What does Thomas know?” but the practical one: “What do others keep asking Thomas?”

Three steps that work without any software:

• For two weeks, pay attention to how often an answer begins with a name. “Ask Thomas.” “Sabine knows.” “Call Mr. Müller.” Note the topic and the person. That list is your first knowledge map.
• Have the answer to a recurring question written down once, with context: which machine, which customer, since when, and why. Not as a wiki project, but wherever the question comes up anyway: in the ticket, the shift handover log, the job folder.
• Start with the two or three people whose list is longest. Not at retirement, but now.

This is exactly where hAiner comes in. Employees ask their question the way they would ask a colleague today. If the answer already exists in a document, a file share, or a system, hAiner finds the passage and cites the source. If the information is missing, that becomes visible too. And when an experienced colleague closes the gap, their answer is kept, linked to the question and its context, instead of vanishing in a one-off conversation. Every resolved question makes the company's memory a little more complete, without Thomas having to maintain a wiki first.

Whoever spots these dependencies early does not have to try to document twenty years of experience in three months when an employee leaves. And saves, as a side effect, the time that disappears every day into searching, asking, and repeated explanations. If you want to know where your company depends on individual people: the knowledge loss check shows you in a few minutes. How to keep those people's knowledge in the company for good is covered on our topic page Securing experiential knowledge.