Verbatim vs Intelligent verbatim transcription: Which is right for research?

The choice between verbatim transcription and intelligent verbatim is an analytical decision about what you can claim based on your data. It needs to be made before fieldwork begins, not after receiving an erroneous transcript. The terms 'clean verbatim’, 'intelligent verbatim', 'edited verbatim' and 'semi-verbatim' are used differently by different providers and sometimes interchangeably.  

What full verbatim transcription captures 

Full verbatim transcription is a complete, unedited record of everything on the recording: 

• Every filler word: um, uh, like, you know 

• Every false start and self-correction 

• Repetitions and stutters 

• Non-verbal cues: [laughter], [long pause], [sighs] 

• Crosstalk and overlapping speech 

• Incomplete sentences left unfinished 

A full verbatim extract looks like this: 

"I... I think we, uh, we tried it once, yeah, and it was... [pause] it was fine, I suppose. I mean, not... not bad." 

Choose full verbatim when how something is said (not just what was said) is analytically significant. Example scenarios: 

• Discourse analysis and conversation analysis, where speech patterns, interruptions, and pauses are the data 

• Interpretive phenomenological analysis (IPA), where hesitation and self-correction reveal meaning-making in real time 

• Healthcare, legal, or compliance research where exact wording carries evidentiary or regulatory weight 

• Any research where a long pause before a sensitive question, or a self-correction mid-sentence, changes the interpretation 

What intelligent verbatim transcription captures 

Intelligent verbatim transcription (also called clean verbatim) removes the fillers, false starts, and repetitions while preserving the speaker's exact meaning and their own words. The same extract in intelligent verbatim loses the hesitation (before "fine") and the self-correction (on "not bad"): 

"I think we tried it once, and it was fine. Not bad." 

Intelligent verbatim is what most researchers mean when they say "transcription" in a commercial research context, and it is the format that most AI transcription tools produce as their default output.

Typically suits scenarios like: 

• Thematic analysis and content analysis, where the goal is meaning extraction, not linguistic pattern analysis 

• Commercial IDI and focus group studies where the deliverable is a client-facing report 

• Studies with high transcript volume, where readability affects how quickly the analysis team can move 

• Any output that non-researcher stakeholders will read directly 

If unsure which your research method requires, start with full verbatim. You can clean a verbatim transcript into intelligent verbatim. However, you cannot recover fillers, pauses, and false starts from an already-cleaned transcript without going back to the audio, which costs time you do not have mid-project. 

Quick comparison 

How flowres.io handles this in practice 

flowres.io generates automated transcripts that sit close to full verbatim by default, capturing what was said with speaker labels and timestamps. The interactive transcript editor then lets researchers clean, edit, and prepare the transcript for analysis without leaving the platform. 

Within the editor: 

• Session & segment naming: Relabel sessions and segments, basis how you want them analyzed eg. Gender, Location only OR Gender, Age, Location, Heavy User 

• Find and replace: Standardise terminology, fix misheard words, or clean filler patterns across the full transcript in one step 

• Speaker name editing: Relabel or standardise speaker identifiers before analysis begins 

• Custom vocabulary: Pre-load specialist terms, brand names, or domain jargon to reduce the correction burden on the editing pass 

For studies requiring human-reviewed transcription at 99% accuracy, human proofreading is available as an add-on. For outsourced transcription across languages or at scale, myTranscriptionPlace handles both full and intelligent verbatim styles across 130+ languages. 

The practical implication: researchers who know they want intelligent verbatim for thematic analysis can use the transcript editor to clean the AI-generated output in one pass, rather than paying for a separate human transcription service from the start. For sessions where full verbatim is required from the outset, the human proofreading add-on delivers a peer-reviewed transcript without the researcher having to change platforms. 

FAQs 

What is the difference between verbatim and intelligent verbatim transcription? 

Full verbatim captures everything spoken, including fillers, false starts, and pauses. Intelligent verbatim removes those elements while preserving the speaker's exact words and meaning. 

Can I convert a full verbatim transcript to intelligent verbatim? 

Yes, and this is the recommended approach when you are uncertain which you need.  

Is intelligent verbatim the same as clean verbatim? 

The terminology is not standardised. Always confirm with your transcription provider exactly what each includes/ excludes, before the project begins. 

Does flowres.io produce verbatim or intelligent verbatim transcripts? 

flowres.io generates automated transcripts which tend to be close to full verbatim, with an interactive editor that lets researchers clean and prepare the transcript themselves. In addition, human proofreading is available as an add-on for sessions requiring verified accuracy.