Section 1
People answer the question you ask, not the one you meant
Ask how something is going and you will get a status report, calibrated to what the person thinks you want and to whether the answer could reflect on them. That is not evasion. It is a reasonable response to a vague question from someone senior. Events work better than evaluations. Ask about the last time something specific happened, walk through it in sequence, and ask what they did next. Concrete recall is harder to smooth over, and it surfaces the workarounds people have stopped registering as workarounds. The pattern of asking, and being seen to act, is what eventually makes people volunteer material unprompted, which [How to Foster a Culture of Storytelling Within Your Organization](/blog/how-to-foster-a-culture-of-storytelling-within-your-organization) treats as a system.
Section 2
Silence, and who is allowed to fill it
Most interviewers talk too much. The instinct after a pause is to offer a suggestion, and the suggestion becomes the frame the person answers within. Waiting through four seconds of silence is uncomfortable and it is where the second, more specific answer usually arrives. Who does the interviewing matters as much as how. A direct manager asking about frustrations will get a filtered account, no matter how good the relationship. Someone one step removed, with no authority over the person's evaluation, gets closer to what happened. Say plainly at the start what the material will be used for and what will not be attributed.
Section 3
Capture verbatim, tag later
Record where consent allows, and keep the person's own phrasing rather than a tidy summary. The specific words a support agent uses for a recurring complaint are usually better than anything marketing would write, and summaries delete exactly that. Where the accounts are going into a shared library, tag them at capture time rather than promising to do it later. Where you need several accounts at once, a structured session is more efficient than sequential interviews, and [Storytelling Workshops for Your Business: How to Start](/blog/storytelling-workshops-for-your-business-how-to-start) sets out how to run one.
Section 4
Running the interviews without distorting the sample
Six to ten conversations of thirty minutes will surface most of the recurring patterns in a team of moderate size. Beyond that you are mainly confirming. Keep the question list to five and let the order move with the conversation. The bigger risk is who you talk to. Interviewing the articulate, available, enthusiastic people produces a flattering and unrepresentative picture. Deliberately include the newest hire, someone in a function you rarely speak with, and if possible a person who left in the last six months. Long-tenured staff have normalised the problems a newcomer can still see, and departing employees have the least reason to manage their answer.
Section 5
What goes wrong after the listening
The most damaging outcome is a round of interviews followed by silence. People remember being asked. If nothing visible follows, the next round gets shorter answers, and the round after that gets scheduling difficulties. The second is breaking the boundary you set, usually with good intentions: a story shared in confidence appearing in a company post because it was too good not to use. That ends candour across the whole organisation, not just with the person involved. The third is confusing what people say with what they do. Interviews tell you how work feels and what people believe; they do not measure frequency or cost. Pair them with observed results, which is what [Before-and-After: ROI Stories from Automated Startups](/blog/before-and-after-roi-stories-from-automated-startups) records.