What actually happens in a workplace investigation interview
A trained investigator is not improvising. The meeting you are about to sit in has a shape, and once you can see the shape it stops being unpredictable.
Across much of the world, investigators are trained in a non-accusatory, information-gathering model. In the UK and in many organizations that adopted the approach internationally, that model is PEACE — preparation and planning, engage and explain, account, closure, evaluation. Three of those five happen outside the room.
You do not need to know the framework to get through the meeting. You need to know one thing about it: the structure is designed to get you talking freely first, and to test what you said second.
Phase one: the opening
The first five to ten minutes. It feels like preamble. It is not.
The investigator introduces everyone present and explains their role. They tell you what the concern is — in enough detail for you to know what you are being asked about. They explain how notes are being taken, whether it is recorded, roughly how long it will take, and what happens afterwards. They may tell you that you can ask for a break, and that you should say if you do not understand a question.
What it is for. Two things, both legitimate. It establishes procedural fairness, which matters to the organization as much as to you — a finding reached in a process where the subject was never told what they were answering falls over the moment it is challenged. And it establishes rapport, because people give fuller and more accurate accounts when they are not frightened. A friendly opening is a professional technique, not a trick.
What to do. Listen to the scope and write it down in their words before you say anything substantive. If they say a concern about expense claims in the second half of last year, that is your boundary, and if the questions later drift outside it you will notice. Ask the practical questions now rather than at minute forty. If you have not been told clearly what the concern is, ask: "Before we start — can you tell me specifically what the concern is that I'm here to respond to?"
The failure mode. Starting your defence here. The opening feels like an accusation, so people begin answering it before any question has been asked. Two things go wrong at once: you stop listening to the information you were about to be given, and you volunteer material into a space the investigator was not going to enter and now has to follow.
Phase two: your account
This is the centre of the meeting and the part almost nobody prepares for correctly.
You will be asked something open. Tell me in your own words what happened. And then the investigator will mostly stop talking. They may say very little for several minutes. They will let silences run.
What it is for. A free account, given without interruption, is the most reliable information an investigator can obtain. Interrupting degrades it — every question narrows what comes next and imports the questioner's assumptions into your memory. So a trained interviewer deliberately does not interrupt. The silences are not there to make you uncomfortable, though they will. People stop talking before they have finished, and a pause usually recovers the rest.
What to do. Describe events in order, starting before the relevant period. Say what you saw, did and said — concrete and first-hand. "I approved the invoice on the 14th after Sam confirmed the work was delivered" is worth more than "I've always been careful about approvals." Distinguish what you know from what you believe, out loud, as you go. Say when you do not know. And when you have finished, stop — what people produce to fill a silence is generally the least considered thing they say all meeting.
The four failure modes
These account for most of the damage people do to themselves.
Arguing instead of describing. The instinct is overwhelming: skip the facts, go to why it was reasonable. But you cannot argue conduct was reasonable before the conduct has been established, and an argument that runs ahead of the facts reads as avoidance of them. Describe first. The reasonableness of what you did is far more persuasive sitting on a clear account of what happened.
Speculating about other people. "I think the real reason this came up is that Alex has had it in for me since the reorganisation." That may even be true. It is still wrong to say here. It is not evidence — you are guessing at someone's motive. It puts you on record attacking a colleague in a document that outlives this process. And it moves the meeting off what happened and onto who is difficult, which is the terrain where you have least control. If there is a genuine issue of bias, raise it separately, factually, with something concrete behind it.
Volunteering scope. You are asked about expenses in the second half of the year, and you mention in passing a thing from two years ago you have always been slightly uneasy about. You have just extended the investigation. It was not in scope; it is now, because the investigator cannot un-hear it and usually has an obligation to follow it. Answer what is asked. That is not evasion — if they ask about it, you answer fully.
Filling gaps with what must have happened. Memory does not store events, it reconstructs them, and under stress it reconstructs them badly. People slide from "I don't specifically remember that meeting" to "I would always have raised it in that situation, so I must have" without noticing they crossed a line. The record may show you didn't. "I don't remember" costs you nothing. A reconstruction that turns out to be wrong can cost you the matter.
Phase three: clarification and challenge
Once you have given your account, the meeting changes character. The investigator goes back through what you said and takes it apart — systematically, usually not aggressively. They break it into topics. They ask you to expand on things you covered quickly. They put documents in front of you. They put other people's accounts to you where those differ. They ask the same thing twice in different words.
What it is for. Detail, because free accounts are accurate but thin. Testing, because your account has to be compared against the record and against other accounts, and divergences have to be explored — that would be done to anyone. And fairness: an investigator is generally expected to put contradictory material to you and give you a chance to respond. If they never put the difficult point to you and it appears in the findings anyway, the process has failed. So when the hard question arrives, it is a sign the process is being run properly.
What to do. Slow down — the gap between question and answer belongs to you. Answer the question actually asked; if you are asked whether you knew about the policy, the answer is yes, no, or I don't remember, not a paragraph about how badly the policy was communicated. Read documents properly before responding, all of it, taking as long as you need. When the record contradicts you, say so plainly: "If the system shows Thursday, then the system is right and I've misremembered." That is a strong answer, not a weak one.
When another account differs, do not attack the person. "I remember it differently. What I recall is X. I can't speak to why Sam remembers it another way." Two people can honestly remember a conversation differently, and saying so is more credible than asserting the other is lying.
The failure mode that matters most here is adjusting your account to fit. A document contradicts you, and rather than saying you misremembered, you find a version that reconciles everything. Then a second document arrives and the version shifts again. An account that moves each time it meets resistance is the pattern investigators are most alert to, and it does more damage than any single inconvenient fact ever could.
Phase four: closure
The investigator summarises what they understood, asks whether that is accurate, and asks whether there is anything else you would like to add. Then they explain what happens next.
This is your last real influence over the record, and people waste it because they are exhausted and want to leave.
Listen to the summary and correct it, specifically. "One correction — you've said I approved it without checking. What I said was that I checked with Sam verbally and then approved it." People let inaccurate summaries pass because correcting feels pedantic after a long meeting. That summary is going into a report.
Use the invitation properly. It is for a point that genuinely got lost, or context nobody asked about. It is not for a closing argument — a speech about the unfairness of the process undoes the impression of someone who described events calmly for an hour.
Ask what happens next. Timeframe. Whether you will see the notes. Whether a further meeting is likely. Who to contact.
And before anything else afterwards — before calling anyone, before going back to your desk — write down what happened. What you were asked, what you answered, what documents appeared, who was present, how long it took. Your recollection of this meeting will degrade within days. If the process runs for months, that same-day record is the only account of it that belongs to you.
The thing to carry in
You are describing events, not defending a case. The people in the room are gathering facts, and the person who decides is somewhere else, reading a document.
Walk in with a dated chronology and you can do the describing calmly, answer what follows without contradicting yourself, and correct the summary at the end because you will notice when it is wrong.
The First Account walks through all four phases in detail, with the questions typically asked in matters like yours and a printable six-page note sheet for capturing the meeting as it happens. Built by someone who ran these interviews for thirty years.