Build a Searchable Research Repository

1. Go to dovetail.com and sign up.
2. Set up your workspace, naming it for your team or product.

What you should see: An empty project space with options to create a new study or import existing data.

1. Start a new project or study inside Dovetail.
2. Upload interview recordings, transcripts, or notes. check the redaction step above before this upload if you haven't already.
3. Let Dovetail run transcription on any raw audio or video before moving to the tagging step.

What you should see: Uploaded files appear in a processing queue, then transition to a transcribed, searchable state. Troubleshooting: Large video files can take a while to transcribe. Start this step early in your setup session rather than waiting until the end.

1. Once transcription finishes, use Dovetail's automatic tagging and theme detection to scan the imported data and apply initial tags.
2. Review the auto-generated tags against a few transcripts you know well, checking whether they group similar comments accurately.
3. Adjust or merge tags that are too narrow or too broad for how your team actually talks about themes.

What you should see: Highlighted passages in each transcript, each attached to one or more tags, with a running tag list building across your uploaded studies. Troubleshooting: If tags feel scattered or overly granular, consolidate similar tags manually early on. a clean tag structure now saves confusion once you have dozens of studies uploaded.

1. Try the AI summarization feature on one completed study to see a synthesized summary of its key findings.
2. Pull a few AI-suggested highlight clips and check them against the full recording to confirm they capture what you'd have chosen manually.

What you should see: A generated summary organized by theme, plus a set of highlight clips tied to specific tags.

1. Once you have more than one study uploaded, use the search or AI chat feature to ask a cross-study question.
2. Confirm any returned answer against the actual clip or quote it cites before repeating it in a stakeholder conversation.

What you should see: A cited answer that references the specific participant, study, and quote it drew from, not just a generic summary. Troubleshooting: If search results feel thin, you likely need more studies uploaded and tagged before cross-study patterns become visible. this tool gets more useful as your repository grows, not on day one.