The State of Scoping and Transcript Review in 2026: Drafts, Audio, AI, and Who Decides

Every transcript that leaves a reporting firm has been read by more than one set of eyes. The reporter writes it and edits it. A scopist may work through it against the steno notes. A proofreader may read the finished pages. The witness may read it and sign. Then one person signs the certificate, and that person answers for every line.

In 2026, the tools around that chain are changing faster than the chain itself. Audio sync is ordinary in CAT software, automatic speech recognition is in the room whether or not anyone invited it, and the profession’s associations and courts have spent this year saying where the lines are. This note covers who does what in review, the difference between draft and certified text, where audio fits, and what the current guidance on AI actually says.

The people in the review chain

NCRA describes the scopist plainly: “A scopist is a professional transcript editor for court reporters.”¹ What sets the scopist apart from an editor or a proofreader, in NCRA’s account, is the ability to compare the reporter’s shorthand to the transcript, so that “mistranslate errors can be identified, thereby helping the court reporter preserve an accurate record.”¹ Stenograph’s description of its scopist edition of CATalyst lists the work in more detail: correcting formatting, punctuation, grammar, and typos; making sure names and terms are spelled correctly; and flagging questionable areas for the reporter.²

The flag matters. A scopist who is unsure does not decide; they flag the spot for the reporter. That habit is the core of good review.

At the end of the chain is the certificate. Under Federal Rule of Civil Procedure 30(f)(1), “the officer must certify in writing that the witness was duly sworn and that the deposition accurately records the witness’s testimony.”³ California’s Code of Civil Procedure section 2025.540(a) asks the same of the deposition officer.⁴ Between the reporter’s edit and that certificate sits the witness’s own review: under Rule 30(e), a deponent who asks before the deposition ends has 30 days to review the transcript and sign a statement listing changes in form or substance and the reasons for them, and the officer notes in the certificate whether review was requested and attaches the changes.³ For a printable sheet that lists a witness’s changes by page and line, there is a free errata sheet template.

Scopists, proofreaders, the witness, and now software can all suggest. The certificate has one author.

AReporter writes and edits

BScopist flags for the reporter

CProofreader reads the pages

DWitness: 30 days to review

ECertificate: one author

Realtime, rough draft, and certified are three different things

Much of the confusion about review comes from treating three products as stages of one.

The guidance on rough drafts is specific. NCRA’s COPE Guidelines for Professional Practice, Section III, say an uncertified draft “may contain errors, some of which could change the accuracy or meaning of the testimony,” may not be filed with the court, and may only be distributed to ordering parties. Each page should carry a header or footer stating “uncertified draft transcript only,” the draft should not include a completed title page, appearance page, or certificate page, and a disclaimer should go with it as a cover sheet.⁶ California puts the same idea in statute: under section 2025.540(b), a rough draft deposition transcript “may not be certified and may not be used, cited, or transcribed as the certified transcript,” and may not be used to rebut or contradict the certified transcript.⁴

Any review tool that touches a rough draft inherits those rules. Its output is still a rough draft until the reporter finishes the work and signs.

ARealtime: draft, unedited, live

BRough: uncertified on every page

CNo certificate page

DCertified: the officer signs

Audio sync is ordinary; control of the audio is the question

Reading the draft against the audio is not new. Advantage Software’s Eclipse page says “all versions of Total Eclipse feature synchronized audio recording and playback at no extra charge,” and that its MediaLink feature “has been part of Eclipse since 1996.”⁷ Stenograph describes its Case AudioSync Playback as providing “realtime synchronization of audio with text as a transcription aid,” with variable playback speed for faint or garbled speakers.² Click a word, hear the moment it was spoken: both vendors include it in the editing products they sell to reporters and scopists.

The harder questions are about the audio itself. NCRA’s COPE Section IV on backup audio media says the use of a backup recording does not change the reporter’s duties toward the record. It treats a recording the reporter makes at their own discretion, not ordered preserved by law, as the reporter’s work product, says a firm cannot require the reporter to produce it unless a court orders it, and tells reporters to check the applicable laws and rules before creating a backup recording at all.⁶ Before audio goes into any review workflow, the reporter should know whose audio it is, who will hear it, and where it will be stored.

What the 2026 guidance on AI says

NCRA’s Board of Directors adopted a new position statement on AI and automatic speech recognition on February 24, 2026. As reported in the Journal of Court Reporting, NCRA’s stated rationale for it acknowledges these technologies’ “capabilities to be helpful as supplemental tools,” and holds that they cannot “create and preserve the record to the extent that a trained, certified human court reporter or captioner can.”⁸

On July 17, 2026, NCRA published COPE guidance for stenographic reporters and captioners on their own use of AI.⁹ Its main points:

• AI may assist, not create. “AI may be used only to support the stenographic reporter’s work and not to generate the record.” Permissible uses “could include technical research and/or clarification of spelling and vocabulary,” and AI “may not be used to create transcripts or captions in place of stenographic capture.”¹⁰
• Quality control, under the reporter. The guidance’s policy statement limits AI to “a limited assistive tool for the stenographic reporter’s sole use in quality control.”¹⁰
• Disclosure. Reporters should be truthful about their own use of AI, and disclosure “should affirm that the stenographic reporter created, reviewed, and certified the record.”¹⁰
• Confidentiality. “There must be no unauthorized transcript uploading, data sharing, storage, or reuse of proceeding content.”¹⁰
• Skill. Reporters “must maintain stenographic proficiency independent of AI tools.”¹⁰

A companion document on AI notetakers gives reporters on-the-record language for when a participant’s bot is in the room, including the statement that the certified transcript prepared by the reporter “is the only certified official record.”¹¹

That guidance speaks for stenographic reporters. It is not the whole landscape. The rules on how the record is made vary by jurisdiction. Rule 30(b)(3)(A) allows deposition testimony to be recorded “by audio, audiovisual, or stenographic means” unless the court orders otherwise.³ The federal court reporting program notes that by law, federal court proceedings are recorded by “shorthand, stenotype, stenomask, or electronic sound recording equipment,” at the district judge’s choice.⁵ Florida’s Eighteenth Judicial Circuit, in an administrative order dated January 8, 2026, makes the digital recording the official record except when a court reporter paid with state funds is present, and requires stenographic reporting for capital trials.¹² And some case orders choose steno outright: the Depo-Provera MDL protocol requires a certified court reporter to record depositions stenographically.¹³ For digital reporting, AAERT offers the Certified Electronic Reporter, Certified Electronic Transcriber, and Certified Deposition Reporter credentials.¹⁴

What every method shares is that a person answers for the transcript, and that the hardest words are the ones that need a listener. The profession’s answers are preparation, interrupting to clarify on the record when testimony is unintelligible, and review against the audio by someone accountable for the result.

What good review looks like

Whatever tools sit in the chain, these questions sort helpful review from risky review:

1. Who decides each change? Every suggestion, from a scopist or from software, should be visible, reversible, and accepted or rejected by the reporter. Nothing should change silently.
2. Is uncertainty shown, or hidden? A list of flagged words helps only if the reviewer remembers that an unflagged word is not a verified word. A clean queue is not a finished transcript.
3. Is the draft still labeled a draft? Output from any review pass should carry the uncertified labels COPE describes until the reporter certifies.
4. Where does the audio go? Ask where audio and text are processed, who can access them, how long they are kept, and whether they are used for anything else. Compare the answers with your rules and with NCRA’s confidentiality guidance if you follow it.
5. Can you work without it? If the tool is unavailable on deadline day, the transcript still has to be produced. NCRA’s guidance asks stenographic reporters to keep that ability.

How DepoStack is approaching this

Transcript by DepoStack is Planned. The idea is to help you check a rough draft against its audio, with AI flagging uncertain words for you to review. You would decide every change and remain responsible for the text. Before testing, we will explain which files you can use and how they are handled. You decide whether the app fits your professional guidance and local rules. The reporter certifies. The software never does. Follow the plans on the Transcript by DepoStack page.