Dictating Investigation Interview Summaries That Defend Themselves

The document that gets attacked in an internal investigation is not the interview; it is the summary. The recorded interview, when one exists, is its own record, but the working summary is the document that opposing counsel, the board, or the regulator reads, because it is the compressed version of the witness's account that the file circulates. An attacked summary is attacked for the same two reasons every time: it looks coached, or it looks incomplete. Dictating investigation interview summaries with explicit attribution and timestamps is the method that answers both attacks before they are made. The method is the reason an account recorded in the room is treated differently from the same account reconstructed a week later, and the difference is what the file's credibility rests on.

A coached summary has a recognizable shape. The language is uniform across witnesses, the qualifiers the witness used have disappeared, and the summary adopts the lawyer's voice rather than the witness's. The reader of such a document reaches one conclusion: the attorney improved the testimony before writing it down. The fix is mechanical: record what the witness said with the witness's own hedging, and mark every summarized statement. The summary is not a cleaner version of the interview; it is a map of the interview, and the map must show the terrain, not the road as it was intended. The witness's own uncertainty and emphasis are the terrain, and they are the part the second interview will test rather than the part the report will replace.

An incomplete summary is the other attack. The summary that omits the second topic, the awkward pause, or the witness's refusal is not a summary; it is a selection, and the selection is attributed to the investigator. The completeness defense is documentation: a summary that states the topics covered and flags the ones not covered. The flag is honest, and honesty is what makes the file stand up. The file that shows its own limits is the file that the reviewing lawyer can defend, while the file that presents certainty invites the question of what was left out. The reviewing attorney's first read of an investigation file is a read for the file's own confidence, and a file that states its gaps has already survived the only test that matters.

The summary is dictated when the door closes, not the next morning, because the thirty minutes after the interview is the only window where your memory holds the witness's words rather than your interpretation of them. Memory does not decay linearly in investigations; it compresses. By morning, the summary you produce is the story you told yourself on the drive home, which is a coherent story but not the interview. The door-closing dictation captures the pauses, the corrections, and the follow-ups, and those are the details that the second interview will test. The recording itself is evidence of the method.

There is a practical ritual that makes the timing work. The witness leaves, the door closes, and the investigator dictates before checking email, before writing the follow-up list, before the next call. The ritual takes ten minutes, and it is the only version of the summary worth keeping. The investigator says the date and the time first, then the witness's name, then the topics in order, then the witness's account, then the flags. The timestamp at the head of the note is what makes the note contemporaneous, and contemporaneity is half the defense. The talk-track sequence, date, witness, topics, account, flags, is also the sequence the file's reader follows later, so the contemporaneity and the structure reinforce each other.

The door-closing note is also a self-check. When the investigator dictates from memory and the account comes out internally inconsistent, that inconsistency is a signal that the witness was unclear or the investigator was not listening, and the signal belongs in the file as a flag rather than being smoothed. Investigators who dictate days later produce fluent notes with none of these seams, and fluent notes are the ones that look rehearsed. The seam is evidence of an honest process, and the honest process is what an investigation file actually sells. The practitioner who has preserved the seam can explain it; the practitioner who has smoothed it cannot, and the distinction appears in the first hour of the examination.

Attribution stays clean through two spoken cues that the whole team uses identically: quote and paraphrase. When the witness's words are to be preserved, the investigator says quote, then the witness's sentence, then end quote, and the note records the words verbatim. When the investigator is compressing the account, the cue is paraphrase, and then the compressed statement, with the witness's hedging words carried into the note. The cues do two jobs: they make the note's truth value visible to the next reader, and they force the investigator to decide consciously, in the moment, which version is being created. The two cues also keep the note's edit history honest, since a quote later marked as a paraphrase becomes a visible change rather than a silent replacement.

The quote cue is for the small set of words that matter: the admission, the instruction, the date, the description of the document. The paraphrase cue is for everything else, and the paraphrase must carry the witness's uncertainty explicitly. A witness who says I think it was March, perhaps gets a paraphrase that reads the witness believed the date was March, and the belief is the paraphrase's subject. The clean rule is that a paraphrase never converts a belief into a fact and never converts a hedge into an assertion. The next reader of the file sees which is which on every line.

The cue discipline survives the handoff. The associate who reads the note months later knows that the sentences under the quote marker are the witness's words and that the sentences under the paraphrase marker are the investigator's compression, and the distinction matters to the final report, where a quote that is actually a paraphrase becomes the document's weak point. The team convention is one page: quote means verbatim, paraphrase means compressed with the witness's hedging, and the reviewer checks a sample. The sample check is what keeps the convention from degrading into a habit. The convention survives because it is cheap to verify, and the verification is what makes the summaries usable weeks later when the recall question finally matters most.

The second interview is the audit of the first one. When a witness is interviewed again, the dictated summary of the second session exists alongside the first, and the differences between the two notes are information, not noise. The investigator should track three classes of difference: memory decay, where the second account is less specific than the first; correction, where the witness changes facts; and amplification, where the second account adds detail that pressure brought out. Each class has a different meaning, and the note should label the class aloud when the difference is spotted. The labeled class is the answer to the question the examiner will ask about the second session's difference, and the label is the difference between a documented methodology and a changed story.

Memory decay between sessions is common and mostly benign, but it warns the team that the first note's specificity is the best evidence available, so the first note is the one the report should rely on, and the second session's vagueness should be flagged rather than merged. Corrections are a different matter: a witness who changes a fact between sessions has created a conflict the file must preserve, and the resolution of the conflict belongs in the summary as an explicit sentence, not as a silent update. The silent update is the fastest way to lose a case in the deposition of the investigator.

Amplification is the class that protects the process. Every fact the witness adds under second pressure makes the first note look restrained, and opponents read restraint as coaching. The antidote in the file is the documented methodology: the dates of both interviews, the cue system, the door-closing note. The methodology, once stated, explains the difference without the investigator having to defend it later. The second-interview check is what turns the file from a collection of notes into a process description, and the process description is the part that survives an attack. The process description also protects the investigation's conclusions, since a conclusion drawn from a documented process is one that can be restated under pressure without a note in hand.

The file discipline is decided before the first interview. There is one review file where the unredacted summaries live, and it is protected by privilege and work product until a court orders otherwise. The file's structure makes the position: the interview room is not mentioned in the summaries, the notes are dated and attributed, and the raw recording, when kept, is stored separately. The review file is the only place where the whole record exists, and the discipline of the whole record is what lets counsel choose what to disclose later rather than discovering that the file has already made the choice.

Who gets the minutes is the question that gets answered before the minutes exist. The client designate and the board receive a periodic status report containing the topics covered and the findings in progress, but not the witness-by-witness summaries, because a board member holding a summary is a board member who can be asked about it in a deposition. The attorney who will be examining the witness gets the summary; the business person does not. When in doubt, the rule is the summary stays in the review file, and the conversation happens by phone, because a spoken conversation has no discoverable copy.

At the close of the investigation, the file gets its final pass. The summaries are reviewed against the report, the flags are resolved, and the unredacted file is either retained under the matter's privilege protocol or transferred to the client's archive as directed. The last dictated note of the matter is the disposition: what the file contains, what was disclosed, and the reason for each disclosure. That note is the document an examiner will eventually ask for, and its existence is the reason the file can be defended. A process that was documented from the first door-closing note to the last disposition is a process that leaves nowhere for an attack to land.