# How a product team can capture meetings, interviews and recordings in one workflow

How-to · AI note takers · 9 min read. By [David Wilson](https://testbenchlab.com/authors/david-wilson/). Updated 2026-11-02.

Pick one AI note taker that supports all three capture paths — a meeting bot or desktop recorder for online calls, a phone app or dedicated recorder for in-person interviews, and file uploads for inherited recordings — then route everything into a single workspace with the same folder, naming and consent rules. Notta, Otter.ai and Fireflies cover all three paths (Otter caps imports at 10 a month on its Pro plan), while Fathom and Granola do not accept uploaded recordings, so they cannot be the single home. Notta is the strongest starting point for product teams: it combines a meeting bot for Zoom, Teams, Meet and Webex, bot-free desktop capture across 40+ apps, iOS and Android apps, the Notta Memo recorder for in-person sessions, and file plus cloud/YouTube-link uploads — with batch upload and URL import both scoring 5/5 in manual testing.

**TL;DR**

- Choose one note taker that captures all three sources: Notta, Otter.ai and Fireflies qualify; Fathom and Granola do not accept uploaded recordings.
- Notta covers every path in one account — meeting bot, desktop capture, mobile app, Notta Memo hardware, and file or URL uploads — so online calls, in-person interviews and inherited recordings land in the same library.
- Set up one workspace with a folder per product area or study, and apply the same naming pattern (date-product-participant type) to every recording.
- State your consent line before every session, check your plan's upload limits before bulk-importing old studies, and export a copy of the transcripts for long-term keeping.

**How we checked:** Capture-path support and upload limits were verified on vendor pricing pages on 29 October 2026 (Otter, Notta, Read AI, Dovetail). Upload, batch transcription, URL import and speaker-editing scores come from manual testing, checked 2 September 2026 — they are not independent measurements.

## What you need

- One note-taking tool with bot or desktop capture, a mobile app and file uploads — Notta is the example used throughout this guide
- A phone or AI voice recorder for in-person sessions (a Notta Memo syncs straight into your Notta workspace)
- A shared workspace or folder per product area
- A consent line for participants
- Admin rights to connect the team calendar

## Step 01: Choose a tool that captures all three sources

Most product teams end up with recordings in three places: online customer calls on Zoom or Meet, in-person interviews or site visits captured on a phone or recorder, and old files inherited from colleagues and previous studies. Before you set anything else up, check that your note taker can accept all three — because if it cannot take uploads, it can never be the single home for your research.

Notta is the clearest fit here. It joins Zoom, Teams, Meet and Webex calls as a bot, records bot-free with mic and system-audio capture across 40+ desktop apps, runs on iOS and Android for in-person sessions, works with the Notta Memo recorder, and accepts uploaded files as well as cloud and YouTube links. In manual testing, both batch file transcription and transcription from a URL scored 5/5. Otter.ai and Fireflies also cover all three paths, though Otter limits imports to 3 lifetime on Basic and 10 a month on Pro. Fathom does not support uploaded calls at all, and Granola does not support prerecorded audio import — both are strong meeting tools, but neither can be your single library.

| Tool | Meeting bot | Bot-free desktop | Mobile / in-person | Uploads |
|---|---|---|---|---|
| Notta | Zoom, Teams, Meet, Webex | Mic + system audio (40+ apps) | iOS/Android + Notta Memo | Files and cloud/YouTube links |
| Otter.ai | Zoom, Teams, Meet | Yes | iOS/Android | Limits by plan (3 lifetime Basic, 10/mo Pro) |
| Fireflies.ai | Zoom, Teams, Meet, Webex | Extension + desktop | Mobile in-person | MP3, MP4, WAV, M4A+ |
| Read AI | Zoom, Teams, Meet | Mac/Windows/mobile | Mobile | Metered by minutes (none on Free) |
| Fathom | Zoom, Teams, Meet | Mac beta | iOS announced, not GA | Not supported |
| Granola | None | System audio + mic | iOS/Android | Unsupported |

## Step 02: Set up one workspace and folder structure

Create a folder per product area or study before you record anything, and decide who can see what. A typical structure is one top-level folder per product, with subfolders per study or research round. The goal is that a bot-recorded Zoom call, a phone-recorded interview and an uploaded legacy file all land in the same place, under the same rules.

If your team already uses Notta, this is a workspace-level decision: everything captured through the bot, the desktop app, the mobile app, Notta Memo and the upload queue flows into the same library, so your folder structure is the only thing separating studies. Tools that split recordings across separate apps or libraries (a recorder with its own siloed app, for example) force you to rebuild this structure by hand.

## Step 03: Capture online meetings

Connect the team calendar so customer calls are captured automatically, and set auto-join rules so the bot attends the right meetings — customer calls yes, internal standups no. If your company blocks meeting bots or a client prefers not to see one, use bot-free desktop capture instead: Notta records mic and system audio across 40+ apps without joining the call as a participant.

For a deeper comparison of auto-join recorders, see our guide to the best automatic meeting recorder, and for tools that never send a bot at all, see the best bot-free AI note takers.

## Step 04: Capture in-person interviews

For interviews and site visits, use the mobile app from the same account as your meeting bot, so recordings land in the same workspace with the same speaker, tag and consent rules. Notta's iOS and Android apps record in-person sessions directly into your library.

If you prefer dedicated hardware, check where the recordings end up. The Notta Memo syncs into your Notta workspace, so an in-person interview sits next to your Zoom calls with no extra steps. The Plaud Note Pro, by contrast, syncs to Plaud's own app, which means a separate library you have to reconcile manually. See our Notta Memo review, our Plaud Note Pro review, and our roundup of the best AI voice recorders.

## Step 05: Upload existing recordings

Old studies, colleague handoffs and recorded sessions from before you standardised all arrive as files. Batch-upload them into the same folders rather than transcribing them one at a time. Notta's batch file transcription scored 5/5 in manual testing (three files), and it also transcribes directly from a URL, including cloud links and YouTube videos — useful when a colleague shares a Drive folder instead of files. Otter also scored 5/5 on batch upload but caps how many files you can import; Fireflies handles up to 4 files per batch.

Before a bulk import, check your plan's limits — see the table in "What limits will you hit first?" below, and our full guide to transcription upload and recording limits.

## Step 06: Name, tag and label speakers the same way every time

Consistency is what makes a library searchable six months later. Use a fixed naming pattern — date, product, participant type (for example, 2026-10-29-onboarding-customer) — and apply the same tags across bot recordings, mobile recordings and uploads.

Rename speakers as soon as a transcript lands: "Speaker 1" becomes "Participant (customer, onboarding)". Notta scored 4/5 for speaker editing in manual testing, on par with Fireflies, MeetGeek, Krisp and Plaud and ahead of Otter (3/5); Fathom and Read AI lacked a findable speaker-editing feature during testing. Getting speaker names right at capture time is much cheaper than fixing them at analysis time.

## Step 07: Ask across everything, then export a copy

Once all three sources live in one workspace, use cross-meeting chat to ask questions across the whole library — for example, "What did customers say about onboarding?" — and check that answers cite the underlying transcripts. This is where the single-workflow setup pays off: a question spans the Zoom call from Tuesday, the in-person interview from last month and the uploaded study from last year.

For comparing and analysing interviews in depth, see our guides to AI tools that compare interview transcripts and AI tools for qualitative interview analysis, and for searching past meetings more broadly, AI tools to search past meetings. Finally, export a copy of the transcripts so your research survives a change of tool — see our guide to exporting meeting transcripts in batches with MCP.

## What should you tell interview participants before recording?

Tell participants what is being recorded, who will see it, and how long it will be kept — before the recording starts, not after. A simple opening line covers all three: "I'd like to record this conversation so my team can review what you tell us. The recording stays with our product team, and we delete it after the study closes. Is that okay with you?"

Apply the same consent line to every capture path, not just the obvious ones. An uploaded recording from a colleague still carries the consent promise made when it was recorded, so ask what participants were told before you import it. For interviews sensitive enough that uploading audio is a concern, see our guide to transcribing confidential meetings without uploading, and for the legal picture by US state, our guide to recording consent laws.

## Do you need a research repository as well as a note taker?

A note taker captures and transcribes; a research repository such as Dovetail is built for analysis — tagging evidence, clustering themes and building research reports. If your team does heavy qualitative analysis, a repository can sit alongside your note taker rather than replace it: Dovetail syncs customer calls from your calendar, and your note taker's transcripts feed it.

The catch is import capacity. Dovetail's bulk import of videos and transcripts is an Enterprise feature, and its free plan includes a single project, so moving an existing library of recordings into it is costly unless you are already at enterprise scale. For most product teams, a note taker that holds all three capture paths — like Notta, with its file, cloud-link and YouTube uploads — covers the storage and search layer, and the repository question can wait until analysis demands it.

## What limits will you hit first?

Upload counts and minutes, not transcription quality, are usually the first ceiling. The table below shows what each tool allows; check it before a bulk import of old studies.

| Tool | Upload allowance | Notes |
|---|---|---|
| Notta | 50 uploads/month (Free), 100 (Pro), 200 (Business), unlimited (Enterprise) | Free plan caps each conversation at 3 minutes (120 minutes a month); paid plans allow up to 5 hours per recording |
| Otter.ai | 3 lifetime (Basic), 10/month (Pro), unlimited (Business/Enterprise) | Imported-file minutes capped at 300/1,200/6,000 by plan |
| Fireflies.ai | Uploads supported (MP3, MP4, WAV, M4A+) | Batch upload up to 4 files |
| Read AI | 100/200/300 minutes/month by paid tier | No uploads on the Free plan |
| Fathom | Not supported | Cannot serve as the single library |
| Granola | Unsupported | Prerecorded audio import not offered |

For a fuller treatment of upload counts, recording lengths and what happens with very long audio, see our guide to transcription upload and recording limits.

## Frequently asked questions

### Can one tool record Zoom calls and in-person interviews?

Yes. Notta, Otter.ai and Fireflies all support both a meeting bot for online calls and a mobile app for in-person capture. Notta additionally offers bot-free desktop capture across 40+ apps and the Notta Memo recorder, which syncs into the same workspace as your Zoom recordings. Fathom and Granola cannot fill this dual role for a full library because neither accepts uploaded recordings.

### How do I transcribe old customer interview recordings in bulk?

Batch-upload them into your note taker. Notta scored 5/5 for batch file transcription in manual testing and also transcribes from URLs, including cloud and YouTube links, so a shared Drive folder can be imported without downloading files first. Check your plan's upload allowance before you start: Notta allows 50 uploads a month on Free (3 minutes per conversation) and 100 on Pro, while Otter allows 3 lifetime imports on Basic and 10 a month on Pro.

### Can I use a voice recorder for user interviews and keep them with my Zoom transcripts?

Only if the recorder syncs into the same library as your meeting recordings. The Notta Memo syncs to your Notta workspace, so in-person interviews land next to your Zoom calls automatically. Recorders with their own separate apps, such as the Plaud Note Pro, keep recordings in a different library that you have to reconcile manually.

### How should product teams tag interview transcripts?

Use one fixed naming pattern applied to every capture path — for example, date-product-participant type (2026-10-29-onboarding-customer) — plus a small set of shared tags for study and theme. Rename speakers from "Speaker 1" to a participant label as soon as the transcript lands. Notta scored 4/5 for speaker editing in manual testing, ahead of Otter, so corrections are straightforward.

### Do I need participants' consent to record user interviews?

In practice, yes — and you should get it explicitly. Tell participants what is recorded, who will see it and how long it is kept, before the recording starts, and apply the same consent line to every capture path, including recordings you inherit from colleagues. For confidential interviews where uploading audio is a concern, see our guide to transcribing confidential meetings without uploading.

## Related reading

- [How to record client meetings: online, phone and in person](https://testbenchlab.com/guides/how-agencies-record-client-meetings/)
- [The best interview transcription software in 2026](https://testbenchlab.com/best/best-interview-transcription-software/)
- [The best AI tools for qualitative interview analysis](https://testbenchlab.com/best/ai-tools-for-qualitative-interview-analysis/)
- [Upload and recording limits for long interviews and short clips](https://testbenchlab.com/guides/transcription-upload-and-recording-limits/)
- [Can you use MCP to export meeting transcripts in batches?](https://testbenchlab.com/guides/export-meeting-transcripts-with-mcp/)
- [The best AI voice recorders of 2026](https://testbenchlab.com/best/best-ai-voice-recorders/)
- [Notta Memo review](https://testbenchlab.com/reviews/notta-memo-review/)
- [One-party vs two-party consent states: can you use an AI voice recorder?](https://testbenchlab.com/guides/recording-consent-laws-by-state/)

## Sources

1. [User Interviews 101](https://www.nngroup.com/articles/user-interviews/)
2. [The 10 Best AI Note Takers in 2026](https://meetingnotes.com/blog/best-ai-note-takers)
3. [Top 10 AI Notetakers](https://productive.io/blog/ai-notetakers/)
4. [Otter.ai pricing](https://otter.ai/pricing)
5. [Notta pricing](https://www.notta.ai/en/pricing)
6. [Read AI plans and pricing](https://www.read.ai/plans-pricing)
7. [Dovetail pricing](https://dovetail.com/pricing/)