Resources · Educational

What Is Verbatim Transcription? A Plain-Language Guide

Verbatim transcription captures every word spoken during a proceeding — not a summary, not an edited version, but the complete spoken record, including false starts, corrections, procedural exchanges, and all verbal contributions from all participants.

The definition of verbatim transcription

Verbatim transcription means producing a word-for-word text record of what was said during a spoken proceeding. A verbatim transcript captures everything said, attributed to the speaker who said it, in the order it was said. It includes not just substantive discussion but also procedural exchanges, interruptions, clarifications, and corrections. Verbatim transcription is different from summary transcription (which captures the gist), edited transcription (which removes repeated words, false starts, and filler), and meeting minutes (which are a structured record of decisions, not a complete spoken record). The choice between verbatim and other formats depends on what the organisation needs the record for and what obligations apply.

When a verbatim record matters

Verbatim records are most important in formal, legal, or adjudicative contexts where the exact words used — not just the substance of what was decided — may need to be reviewed later. In a tribunal hearing, a witness's exact testimony may be significant. In a council meeting, the exact wording of a motion before and after an amendment may matter. In an investigation, the precise words an interviewee used in response to a specific question may be critical. In these contexts, a summarised or edited record is not sufficient — not because summaries are inherently inadequate, but because they do not allow verification of exactly what was said. If a question later arises about whether a witness contradicted an earlier statement, or whether a motion was worded as proposed or as amended, the verbatim record is what allows that question to be answered definitively.

What AI verbatim transcription produces — and where review matters

AI transcription systems produce a text that is as close to verbatim as the technology allows — capturing what was said, attributed to speakers, with timestamps throughout. This is not the same as a professionally produced, certified verbatim transcript, but it is a structured verbatim draft that significantly reduces the work required to produce one. For AI verbatim transcription to be reliable, the audio quality needs to be adequate, speakers need to be identifiable, and a reviewer needs to check the output against the audio for accuracy. The verbatim nature of the transcript means that errors — wrong words, missed speaker transitions, misidentified speakers — can affect the record in ways that matter.

Frequently Asked Questions

Does verbatim include filler words like 'um' and 'uh'?
A true verbatim transcript includes all spoken sounds, including filler words, false starts, and verbal corrections. Some organisations prefer a 'clean verbatim' approach that removes filler words while preserving the substance. The appropriate standard depends on the context and the organisation's requirements.
Is verbatim transcription always necessary?
No. For many organisational records, a clear summary or structured record of decisions is sufficient. Verbatim transcription is most appropriate when the exact words used may need to be reviewed — in formal hearings, legal matters, investigations, and adjudicative proceedings.
How accurate is AI verbatim transcription?
Accuracy varies with audio quality, speaker characteristics, acoustic conditions, and subject matter. Well-performing AI systems achieve high accuracy under favourable conditions but can make characteristic errors with names, technical terms, and unclear audio. Human review is important for any verbatim transcript that will be relied upon.