How to run knowledge capture interviews with fleet veterans

Who runs it, and who gets interviewed

The interviewer does not need to be the most senior person available. Often better if they are not. A curious junior technician asking sincere questions gets richer answers than a manager working through a checklist, and the junior is the one who will use the answers. What the interviewer needs is enough shop fluency to follow the story and ask the next question.

Pick interviewees by risk, not by seniority alone. NASA's report suggests asking the team a pointed question: what are you most worried about losing when this person leaves? Whoever the answers cluster around goes first. A quiet specialist whom everyone consults about one stubborn subsystem may matter more than the longest tenured name on the roster.

Hold it in the yard, not the conference room

Memory is anchored to machines. A veteran asked in a meeting room what they know about the older units will shrug and offer generalities. The same veteran standing at the fender of the actual truck will talk for twenty minutes without a pause, because the object in front of them keeps prompting the next recollection. Walk the yard. Stop at the units with reputations. Let the equipment set the agenda.

Keep sessions short and specific

NASA found that short recordings of one specific practice hold up better than long career retrospectives, and shop experience agrees. A focused session runs 30 to 60 minutes and covers one theme: one problem unit, one subsystem, one seasonal ritual. Long sessions produce tired answers and unusable recordings. A series of short ones, spread over months, builds a real archive without stealing whole days from the schedule.

Questions that actually work

Open with specifics, never with "tell me everything you know." Questions that reliably produce usable knowledge:

• Which unit in this fleet would you never lend to another yard without a warning label, and what is the warning?
• Where does the manufacturer schedule get it wrong for our routes and our weather?
• What failure did you see coming last time that nobody else saw, and what tipped you off?
• What do you check first when this model comes in with a complaint you cannot reproduce?
• Who do you call when you need a part fast, and what do you say?
• What rule do we follow that nobody remembers the reason for? Do you remember the reason?

Then follow the thread. The word "why" is the whole technique. Every override the veteran mentions has a story behind it, usually a failure, and the story is the part that makes the knowledge transferable. A rule without its reason becomes superstition in the next generation's hands. More on that in what leaves when a veteran mechanic retires.

Record voice, then transcribe

Do not make the veteran write, and do not make the interviewer scribble. Record the conversation, with permission, and transcribe it afterward. Voice captures the hedges and conditions that matter: "usually," "only in winter," "not on the newer ones." Those qualifiers are data. A phone in a shirt pocket is equipment enough.

Process while it is warm

A raw transcript is compost, not a knowledge base. Within a week, break each session into discrete entries: one tip, one override, one warning per entry, each tagged with the equipment it applies to and the conditions under which it holds. Have a supervisor or a second veteran read each entry before it becomes official, because a few veteran habits are workarounds for problems that no longer exist, and review is what separates lore from fact. Then put the entries somewhere the whole team can search, because knowledge filed in a binder is knowledge lost on a slower schedule. The storage question is covered in the pillar guide to preserving institutional fleet knowledge.

Cadence beats heroics

One marathon interview the week before retirement recovers scraps. One short session a month, started two years out, recovers a career. Put a recurring hour on the calendar and treat it like a PM interval: unglamorous, scheduled, and cheaper than the failure it prevents.

How YardWise fits

YardWise gives the output of these interviews somewhere useful to live. Veterans record an entry by voice, the system transcribes it, entries get tagged to equipment and conditions and routed to a supervisor for approval, and any technician can then ask a question in plain English and get the local note alongside the manufacturer spec. See how it works. The interviews are worth running either way.