{"site":{"name":"Koji","description":"AI-native customer research platform that helps teams conduct, analyze, and synthesize customer interviews at scale.","url":"https://www.koji.so","contentTypes":["blog","documentation"],"lastUpdated":"2026-11-08T22:03:05.973Z"},"content":[{"type":"blog","id":"a36d35ee-422e-401b-8d04-8b634306992d","slug":"transcript-redaction-tools-compared-2026","title":"Transcript Redaction in Research Tools, Compared (2026): What Each Removes and Where","url":"https://www.koji.so/blog/transcript-redaction-tools-compared-2026","summary":"There are four main ways to remove personal details from user research interview transcripts in 2026. Manual scrubbing is free but misses repeats and leaves the original and recording in place. Dovetail and Marvin redact on import on their Enterprise plans, with video blurring or masking. Speech APIs such as AssemblyAI and Azure AI Language redact entity types you choose in code. Koji's Remove personal details setting, on every plan including Free, replaces names, emails, phone numbers, addresses and employers with labels when each interview finishes, before analysis, overwrites the original and drops the recording if anything was removed, while keeping products, job titles and cities. None of them anonymise data: quasi-identifiers such as role, company size and city still need a human check.","content":"If you need names, emails and employers out of interview transcripts before other people read them, you have four realistic routes in 2026: scrub by hand, let your research repository redact on import (Dovetail and Marvin both do this on their Enterprise plans), call a speech-to-text API with redaction switched on (AssemblyAI, Azure AI Language), or have the interview tool clean each transcript the moment the interview ends (Koji's Remove personal details setting, on every plan including Free). They differ on four things that matter more than the feature list: **what gets removed, when it runs, whether the original and the recording survive, and who can switch it on.**\n\n**Try it now:** in Koji, open a study, go to **Branding & Settings**, open **Privacy** and turn on **Remove personal details**. Every interview that finishes after that is cleaned before any analysis runs.\n\nNone of the four makes a transcript anonymous. They remove direct identifiers. A job title, a city and a company size can still point to one person, which is why the last section of this post is a checklist, not a promise.\n\n## Why this matters more as you move upmarket\n\nPersonal details are the most common thing lost in a breach. IBM's *Cost of a Data Breach Report 2025* found customer personal data was involved in **53% of breaches**, and put the global average cost of a breach at **$4.44 million**. Interview transcripts are full of exactly that data: people say their own name, their manager's name and their employer without being asked.\n\n![A buyer and a vendor team in a meeting room going through security questions about how customer data is handled](https://images.unsplash.com/photo-1573164574511-73c773193279?crop=entropy&cs=tinysrgb&fit=max&fm=jpg&ixid=M3w5NDUzMDN8MHwxfHNlYXJjaHw1fHxidXNpbmVzcyUyMG1lZXRpbmclMjBjb25mZXJlbmNlJTIwcm9vbSUyMGxhcHRvcHxlbnwxfDB8fHwxNzk0MTIwNjEzfDA&ixlib=rb-4.1.0&q=80&w=1080)\n\n_Photo by [Christina @ wocintechchat.com M](https://unsplash.com/@wocintechchat?utm_source=koji&utm_medium=referral) on [Unsplash](https://unsplash.com/?utm_source=koji&utm_medium=referral)_\n\nThe other pressure is the security review. The first enterprise customer who asks \"how do you handle interview data?\" wants to hear where personal data is removed and who can see the raw version. \"We are careful in Slack\" is not an answer that survives a procurement questionnaire.\n\nAnd removing names is not the same as anonymising. Latanya Sweeney showed in 2000 that **87% of the US population** could be uniquely identified by just ZIP code, gender and date of birth. Rocher, Hendrickx and de Montjoye (*Nature Communications*, 2019) estimated that **99.98% of Americans** could be correctly re-identified in any dataset using 15 demographic attributes. Redaction handles the names. You still have to handle the combinations. Our guide to [quasi-identifiers in research data](/docs/quasi-identifiers-research-data-reidentification) goes deeper on that part.\n\n## The four routes, compared\n\n![Table comparing manual scrubbing, Dovetail, Marvin, speech APIs and Koji by what each removes, when it runs and what happens to the original](/visuals/transcript-redaction-routes-compared/transcript-redaction-routes-compared.png \"The question that separates them is whether the original and the recording still exist afterwards\")\n\n### 1. Manual scrubbing in a document\n\n**Who it is for:** a team running a handful of interviews a quarter, sharing a few quotes at a time.\n\nSomeone opens the transcript, finds names, emails and employers, and replaces them before pasting quotes into a deck or a channel. It costs nothing and you control every decision.\n\nThe weaknesses are the obvious ones. It depends on the person doing it being thorough every single time, it is easy to miss a name repeated twenty minutes later, and the raw transcript and the recording still sit wherever they were saved. It also does not scale: by the time you are running interviews every week, nobody is scrubbing every transcript before it lands in a shared folder.\n\n### 2. Research repositories: Dovetail and Marvin\n\n**Who it is for:** teams that already centralise interviews from many sources in one repository and are on an enterprise contract.\n\n**Dovetail** offers manual and automatic redaction, and its help centre says both are available only on the Enterprise plan. Workspace admins set it up under Settings, Compliance, Data redaction. By default it targets common personal data such as names, addresses, email addresses, phone numbers and bank details, and admins can adjust what it looks for. Each item is screened as it is imported, and the setting applies to data imported after it is switched on; for earlier data, Dovetail says to contact support. Its \"magic redact\" can also blur and mute sections of uploaded video and redact the matching transcript, and redactions carry through to highlights that reference them. Dovetail is clear that automatic redaction is AI-driven and may miss things, and recommends reviewing it. It has also published a changelog entry for restoring redacted data, so ask whether the original is retained if that matters for your policy.\n\n**Marvin** can detect and mask personal information in transcripts, audio and video recordings, and separately in survey responses and support tickets; its help centre lists that second capability as Enterprise only. You turn it on per file rather than for the whole workspace. It covers names, email addresses and banking information, and also sensitive categories such as nationality, race and gender. Marvin notes that it currently masks personal information in English only, that processing can take from a few minutes to a few hours depending on the recording, and that admins can keep an exclusion list of words that should never be masked.\n\nThe strength of this route is breadth: it works on anything you import, including recordings from Zoom or a moderated session. The trade-off is that the raw data reaches the repository first and is cleaned on arrival, and the feature sits on the top plan.\n\n### 3. Speech-to-text APIs: AssemblyAI and Azure AI Language\n\n**Who it is for:** teams with developers who already run their own transcription pipeline.\n\n**AssemblyAI** lets you choose which entity types to redact (person names, phone numbers, email addresses, card numbers, dates of birth and many more), and replace them either with hash characters or with the entity type in brackets. It can also produce a redacted audio file with the matches bleeped or silenced; its docs say that file is available for 24 hours.\n\n**Azure AI Language** has a conversational PII feature that finds personal details in transcripts and chats across predefined categories, returns the redacted text, and for transcripts gives the audio timings of each detail so you can cut it from the recording. Microsoft's docs say the generally available conversational model supports English, and that async results are purged after 24 hours.\n\nThese are precise and configurable, and built for call centres and developers. Someone has to write and maintain the code, decide which entities matter for research (an API does not know that \"Notion\" is a finding and \"Northwind\" is an employer), and wire the output into wherever the team reads interviews.\n\n### 4. Redaction inside the interview tool: Koji\n\n**Who it is for:** teams running AI-moderated interviews who want transcripts clean before anyone, including the analysis, reads them.\n\nKoji's **Remove personal details** is a per-study setting. When it is on, each interview is cleaned as soon as it finishes and before any analysis runs:\n\n- People's names, email addresses, phone numbers, street addresses and postcodes, and the organisations people work for or are clients of, are replaced with labels: **[NAME], [EMAIL], [PHONE], [ADDRESS], [COMPANY]**.\n- The same details are removed where the AI interviewer repeated them back.\n- Products, tools and brands people mention stay, as do job titles, cities and countries, so \"we moved from Trello to Notion\" keeps its meaning.\n- The original transcript is overwritten, not hidden. If anything was removed from a voice interview, the recording is not kept either.\n- Summaries, per-question results, report quotes and exports are all built from the cleaned text.\n\nIt is available on every plan, including Free, and you can also switch it on by asking Chat with Koji or an AI assistant connected over MCP.\n\n![Koji report summary with a side panel of supporting participant quotes behind one finding](/email/shots/shot-evidence-panel.jpg \"Report quotes are drawn from the cleaned transcript, so a quote you paste into a deck carries labels, not names\")\n\nThe limits, stated plainly: it only cleans interviews that finish after you turn it on, it leaves pre-interview form answers as they are (those are fields you asked for by name), and because it is automatic it can miss a detail. You cannot get the original back, by design, so turn it on only for studies where you will never need names. If you need to follow up with someone, collect their email in the pre-interview form rather than relying on what they say in the interview.\n\n## Which route should you pick?\n\n![Flowchart for choosing a transcript redaction route, from whether you need names to checking quotes before you share themes](/visuals/choose-transcript-redaction-route-flowchart/choose-transcript-redaction-route-flowchart.png \"Most teams end up combining two routes: cleaning at the source and a human check before sharing\")\n\nA few rules of thumb:\n\n- **You need names later** (to follow up, to pay incentives, to link to CRM records): keep the originals in a restricted place and share only scrubbed excerpts. Automatic removal is the wrong tool.\n- **You run interviews in Koji:** turn on Remove personal details before the first interview. Cleaning at the source means the raw version never reaches your report, your exports or your teammates.\n- **You import recordings from many places** and already pay for Dovetail or Marvin Enterprise: use their redaction on import, and check what happens to the original.\n- **You run your own transcription pipeline:** use an API, and spend time on the entity list. Employers and clients are the details that most often identify a B2B participant.\n\n## What none of them handle\n\nEvery route above removes direct identifiers. None of them reliably removes:\n\n- **Combinations.** \"Head of RevOps at a 40-person fintech in Lisbon\" is one person.\n- **Unique stories.** \"The week our CEO resigned on stage\" needs no name.\n- **Things said in other ways.** A nickname, a misspelt surname, \"my boss, the one who joined from Google\".\n- **Data collected elsewhere.** Form answers, calendar invites, incentive records and your CRM still hold the names.\n\nThat is why each tool's own documentation says to review results before sharing. The [anonymising interview data guide](/docs/anonymizing-customer-interview-data) covers techniques such as generalising roles and bands, and [k-anonymity for segment reporting](/docs/k-anonymity-segment-reporting-minimum-base-size) covers when a segment is too small to report.\n\n## Who sees your transcripts, and what to check before you share\n\nRedaction decides what is in the transcript. Access decides who reads it. Before a quote leaves the research team, run through this list.\n\n![A small team going through printed interview notes at a table before deciding which quotes to share more widely](https://images.unsplash.com/photo-1616587656977-ac36a5a430bc?crop=entropy&cs=tinysrgb&fit=max&fm=jpg&ixid=M3w5NDUzMDN8MHwxfHNlYXJjaHw0fHx0ZWFtJTIwbWVldGluZyUyMHJldmlld2luZyUyMHByaW50ZWQlMjBkb2N1bWVudHN8ZW58MXwwfHx8MTc5NDEyMDYxM3ww&ixlib=rb-4.1.0&q=80&w=1080)\n\n_Photo by [LinkedIn Sales Solutions](https://unsplash.com/@linkedinsalesnavigator?utm_source=koji&utm_medium=referral) on [Unsplash](https://unsplash.com/?utm_source=koji&utm_medium=referral)_\n\n![Checklist of eight things to check before sharing redacted interview data, from scanning quotes for missed names to setting a deletion date](/visuals/redacted-interview-data-sharing-checklist/redacted-interview-data-sharing-checklist.png \"Eight checks that catch what automatic redaction misses\")\n\nIn practice:\n\n1. Scan quotes for names, nicknames and emails the tool missed.\n2. Look for rare mixes of role, company size and city.\n3. Remove stories only one person could tell.\n4. Check what the pre-interview form collected.\n5. Share themes and short quotes, not full transcripts.\n6. Give each audience the least access it needs.\n7. Set a date to delete raw data you no longer need.\n8. Tell participants how their details are removed. Our [consent form wording guide](/docs/consent-form-wording-disclosure-research) has language you can reuse.\n\nIf you are answering a customer's security questionnaire, the [AI interview data privacy and security guide](/docs/ai-interview-data-privacy-security) lists the questions buyers ask and what a good answer covers, and the [GDPR guide for AI user research](/docs/gdpr-compliant-ai-user-research) covers the legal side.\n\n## Common mistakes\n\n- **Turning redaction on after the study launched.** In Koji, Dovetail and Marvin alike, the setting applies to what comes in afterwards. Switch it on before the first interview.\n- **Calling cleaned transcripts \"anonymous\".** Say \"personal details removed\". It is accurate, and it is what your consent form should promise.\n- **Forgetting the recording.** A redacted transcript next to the original audio has not removed anything.\n- **Removing too much.** A tool that strips product and competitor names destroys the finding. Check that brands and tools survive.\n\n## Start with your next study\n\nIf your interviews already run in Koji, the cheapest route is the one at the source: open the study, go to **Branding & Settings**, then **Privacy**, and turn on **Remove personal details** before you send the link. Then use the checklist above before every share. If you are new to Koji, [start a free study](https://www.koji.so) and the setting is there from day one. For the method behind it, read our deeper guide to [automatic PII redaction in user research](/blog/pii-redaction-user-research).\n\n**Sources:** IBM, *Cost of a Data Breach Report 2025*. Latanya Sweeney, *Simple Demographics Often Identify People Uniquely*, Carnegie Mellon University, 2000. Luc Rocher, Julien Hendrickx and Yves-Alexandre de Montjoye, \"Estimating the success of re-identifications in incomplete datasets using generative models\", *Nature Communications*, 2019. Dovetail help centre, \"Blur and redact\" and \"Manage PII in Dovetail\". Marvin help centre, PII masking articles. AssemblyAI docs, \"Redact PII from transcripts\". Microsoft Learn, \"Detect and redact PII in conversations\".\n","category":"Comparisons","lastModified":"2026-11-08T07:20:15.136195+00:00","metaTitle":"Transcript Redaction Tools for User Research, Compared (2026)","metaDescription":"How Dovetail, Marvin, AssemblyAI, Azure and Koji redact interview transcripts: what each removes, when it runs, what happens to the original, and the plan it needs.","keywords":["transcript redaction tools","pii redaction research tool","anonymize interview transcripts tool","dovetail redaction","marvin pii masking","assemblyai pii redaction","remove names from interview transcripts"],"aiSummary":"There are four main ways to remove personal details from user research interview transcripts in 2026. Manual scrubbing is free but misses repeats and leaves the original and recording in place. Dovetail and Marvin redact on import on their Enterprise plans, with video blurring or masking. Speech APIs such as AssemblyAI and Azure AI Language redact entity types you choose in code. Koji's Remove personal details setting, on every plan including Free, replaces names, emails, phone numbers, addresses and employers with labels when each interview finishes, before analysis, overwrites the original and drops the recording if anything was removed, while keeping products, job titles and cities. None of them anonymise data: quasi-identifiers such as role, company size and city still need a human check.","aiKeywords":["transcript redaction","PII redaction user research","Dovetail redaction","Marvin PII masking","AssemblyAI redaction","Azure conversational PII","Koji remove personal details","quasi-identifiers"],"aiContentType":"guide","faqItems":[{"answer":"Dovetail and Marvin redact on import on their Enterprise plans, AssemblyAI and Azure AI Language offer redaction through their speech and language APIs, and Koji removes names, emails, phone numbers, addresses and employers from each interview when it finishes, on every plan including Free.","question":"Which research tools redact personal details from interview transcripts?"},{"answer":"No. Redaction removes direct identifiers such as names and emails. Combinations of role, company size and city, or a story only one person could tell, can still identify someone, so review quotes before sharing.","question":"Does redaction make an interview transcript anonymous?"},{"answer":"No. Koji overwrites the original transcript on purpose, and if anything was removed from a voice interview the recording is not kept either. Collect any details you need later, such as an email for follow-up, in the pre-interview form.","question":"Can I get the original transcript back after Koji removes personal details?"},{"answer":"No. Products, tools and brands people mention stay, as do job titles, cities and countries, so findings keep their meaning. Only people, contact details, addresses and the organisations people work for or are clients of are replaced.","question":"Does Koji remove product and competitor names?"},{"answer":"In Koji it applies to interviews that finish after you turn it on. Dovetail also applies it to data imported afterwards and asks you to contact support for earlier data. Turn it on before the first interview.","question":"Does turning on redaction clean interviews I already ran?"},{"answer":"It works if you have developers and your own transcription pipeline. You choose the entity types in code, and you have to decide which ones matter for research, since an API cannot tell a tool someone uses from the company they work for.","question":"Is a speech API a good way to redact research interviews?"}],"relatedTopics":["PII redaction","research data privacy","anonymising interview data","research repositories","security reviews"]}],"pagination":{"total":1,"returned":1,"offset":0}}