Your best senior analyst is retiring next year. You know it is coming. You have started thinking about transition plans, documentation sprints, maybe some shadowing sessions for her replacement.
Here is what you are going to capture: her process steps, her file locations, her vendor contacts, her recurring deadlines.
Here is what you are going to lose: the judgment calls. The "I just know when something's off." The workaround she invented six years ago because the system can't handle a specific edge case. The way she reads a report differently in Q4 because seasonal patterns shift the numbers. The three questions she asks in every stakeholder meeting that prevent problems nobody else sees coming.
That second list is worth more than the first. And almost none of it is written down.
The gap between the SOP and reality
Every organization has documented processes. And every organization has experienced workers who quietly deviate from those documents in dozens of small ways that make the actual work succeed.
Tacit knowledge is the gap between what the SOP says and what actually happens. Decision rules nobody formalized. Quality checks experts run intuitively. Risk signals that trigger a change in approach. Workarounds for steps that don't work as documented.
This gap is not a failure of documentation. It is a feature of how expertise works. As people get better, their knowledge goes underground. It becomes automatic, invisible, and inaccessible to anyone who hasn't spent years building the same instincts.
Ask about the hard days
The standard approach to knowledge transfer does not work. You sit someone down and ask "what do you do?" They give you a sanitized, simplified version that matches the documentation.
Better question: "Tell me about a time when it almost went wrong."
Stories surface what procedures miss. When people talk about near-misses and difficult days, they reveal the invisible decisions they make. They describe the signals they watch for, the patterns they have learned to recognize, the moments where the documented process breaks down and experience takes over.
One 30-minute conversation about a hard day will tell you more about what someone actually knows than a week of process documentation.
One problem, not three
Here is what most organizations do not see: the methods that capture tacit knowledge for succession planning are the same methods that improve processes, and the same methods that prepare you for AI.
Knowledge continuity: surface the invisible decision rules before experienced people leave. Process improvement: understand how work actually happens (not how it is documented) before you can improve it. AI readiness: articulate expertise in explicit terms before AI can apply it.
Three business cases. One set of methods. Organizations that treat these as separate initiatives end up doing the same foundational work three times, or (more commonly) never doing it because each initiative alone does not seem worth the investment.
Have you ever seen an AI pilot fail because the team could not articulate the expertise the AI was supposed to replicate? That is not an AI problem. It is a knowledge extraction problem.
The bottom line
Your succession planning problem, your process improvement problem, and your AI readiness problem are the same problem. They all require surfacing the expertise your experienced people carry but can't articulate. Solve it once, and you solve all three.