When Tribal Knowledge Quietly Becomes IT Policy
Every IT team has at least one rule that nobody wrote down. New accounts get configured a particular way because the longest-serving technician says so. A laptop setup step happens on certain days because that is how it worked under the old system. A newer colleague follows the instruction once, discovers it conflicts with how the identity platform actually provisions accounts, and causes exactly the problem the rule was supposed to prevent. When they ask why the rule exists, the answer is: that is how we have always done it.
This is tribal knowledge turning into policy. It is rarely malicious, and the veteran is often right more than they are wrong. But a rule that cannot be checked against a record will eventually outlive the reason it was created, and the team will keep enforcing it anyway.
What is tribal knowledge in IT, and why is it a problem?
Tribal knowledge is the know-how that lives in people's heads and habits rather than in the documentation. In IT, it covers the workaround everyone uses for a temperamental system, the exception agreed with a department years ago, the step a senior engineer always adds to a procedure. Much of it is valuable. Some of it is the only reason certain things work at all.
The problem starts when it quietly overrides the documented process. Systems change underneath habits. An identity platform starts provisioning accounts automatically on a schedule. A device management tool takes over a task that used to be manual. A security review changes what is permitted. The documentation may even be updated. But if the habit is enforced by someone with authority and tenure, and nobody can easily point to the record, the habit wins.
Why does undocumented practice beat the documented process?
Because the person who remembers is easier to reach than the record. To challenge an established habit, a newer team member has to find the current documented process, the change ticket that altered it, and ideally the conversation where the reason was discussed. Those live in different places: a wiki page, a ticketing system, an email thread, a chat channel. Finding them takes an afternoon. Asking the veteran takes thirty seconds, and the answer comes with confidence.
Research on workplace knowledge shows how common this dependency is. Panopto's Workplace Knowledge and Productivity Report found that 60 percent of employees said it was difficult, very difficult, or nearly impossible to get vital information from colleagues, and that knowledge workers lost an average of 5.3 hours a week either waiting for information from coworkers or recreating knowledge that already existed. When getting the real answer is that slow, people stop checking.
What does it cost when habits override process?
The visible cost is rework and inconsistency. Two technicians handle the same request in different ways. A new employee's first day goes wrong because a step was done on the wrong schedule. A change that the documented process would have caught slips through because someone followed the remembered version instead. Each incident is small. Together they shape how reliable the IT department looks to the rest of the business, especially during employee onboarding, when IT makes its first impression on every new hire.
The less visible cost is risk. Undocumented practices are, by definition, unreviewed. A habit that made sense a decade ago may now conflict with security policy or audit requirements, and nobody notices until something goes wrong. It also concentrates authority in tenure rather than evidence, which makes it harder for newer staff to raise legitimate concerns and harder for managers to know which way of doing things is actually correct.
How do you find out why an IT process works the way it does?
The answer is almost never a new documentation initiative. Most teams already have the record; it is just scattered. The original process is in a wiki. The change that altered it is in a ticket. The reason is in a chat thread or an email chain with the department that asked for the exception. What is missing is a way to search all of it at once.
Retrieval-augmented generation, or RAG, does exactly that. It indexes content where it already lives, retrieves passages that match the meaning of a question rather than its exact wording, and produces an answer with citations and dates. A technician can ask "when are new hire accounts enabled, and has that changed" and get the current documented process, the change ticket that introduced automatic provisioning, and the related discussion, each linked. The conversation with the veteran stops being a contest of memory and becomes a review of the record.
How does RetrieveIT help IT teams replace memory with records?
RetrieveIT connects to the systems where IT and help desk teams keep their knowledge, including Confluence, SharePoint, Jira, Microsoft Teams, Outlook, and Google Drive, and makes them searchable together from one box. Every answer cites the source system, document, author, and date, so it is clear whether a practice is documented, when it last changed, and who changed it.
Permissions from each source are enforced at query time, so team members only see what they are already entitled to see. New content syncs within minutes, so when a process is updated, the answers update with it. The same applies to processes that cross departments, like the start-date handoffs between HR teams and IT that new-hire provisioning depends on.
Tribal knowledge is not the enemy. The goal is to make sure that when a habit and a record disagree, anyone on the team can find the record in seconds.
Replace "we have always done it this way"
Connect Confluence, Jira, SharePoint, and Teams to RetrieveIT and give your team cited answers about how and why your processes work, free for 14 days. No credit card required.
Get Started Free