What "certified" actually promises
The certificate at the back of a transcript is a named person's attestation that the words are the testimony. That is the whole guarantee: not that the recording was flawless, but that a human read the record end to end and stands behind it. It matters most when the transcript gets challenged — an errata fight is about what the witness actually said, and the answer rests on the audio and the reporter's account of making the record.
Errata exist because memory is imperfect
Every jurisdiction gives a witness a window to fix errors in what the transcript says they said — wrong dates, switched names, a word the audio does not support. Working the other direction is harder: practitioners keep describing transcripts that arrive with pervasive errors and the line-by-line recovery burden that follows. Some of that is the writer, some is the audio, and some is the difference between what a witness said and what got captured. The correction process is the repair path, and it is slower when the problem is structural rather than a stray date.
What a correction can and cannot fix
A correction fixes the words under the reporter's signature, through the state's process and within its window — it does not retroactively improve a bad recording. That is why the quality pipeline matters before delivery: the reporter's realtime notes, the scopist's read against the audio, the proofread, and the certification. Every stage is a chance to catch a mistranslation before it ships instead of after. And it is why uncertified output says so on every page — a rough draft is for preparation, never the record.
Questions worth asking an agency
Ask who reads the transcript before you do, how corrections are handled when the witness signs them, and what happens when the notes and the audio disagree. If the answers are vague, the pipeline is vague. Our turnaround tiers describe how the read-through fits each delivery speed; request a quote and you get the same answers with a date attached.