Set five export options once. Every DSAR export after that is drag, drop, done — Teams transcripts, mailboxes, and SharePoint files included.
Microsoft Purview eDiscovery is where most organizations already gather the documents a data subject access request covers: mailboxes, Teams chats, SharePoint and OneDrive files. What Purview hands you at the end is an export package — and redacting third-party personal data out of that package is the slow, manual part of every DSAR.
SafeRedact is built to consume that package as-is. With the right export settings, the ZIP that Purview produces can be dropped straight into SafeRedact Enterprise: Teams conversation transcripts are parsed natively, messages are processed individually, folder structure survives into your redaction log, and files that don't apply (load files, logs) are ignored automatically.
In the Purview portal, open your eDiscovery case, run the search scoped to the data subject, and select Export. Then set these options:
| Export option | Set to | Why |
|---|---|---|
| Mailbox export format | Individual .msg files (not PST) | Every message becomes a directly processable file — no container limits, no extraction step. |
| Teams and Viva Engage conversations | Enabled | Chats arrive as threaded HTML transcripts, which SafeRedact reads natively — participants, timestamps, and metadata table included. |
| Include folder and path of the source | Enabled | Original folder structure survives into document names, so your redaction log stays human-readable. |
| Give each item a friendly name | Enabled | Files are named for what they are, not a GUID. |
| SharePoint / OneDrive documents | Native files | PDF, Word, and Excel are all first-class formats in SafeRedact. |
Download the export when it completes. Large exports arrive as several ZIP files — that's fine. You'll drop them all in.
If you cull in a review set first and export from there, the same settings apply where offered. Two extras appear in Premium exports: an optional text files layer and a load file for litigation platforms. Skip the text-file option — SafeRedact extracts its own text from the natives — and don't worry about the load file: unsupported file types inside the ZIP are skipped silently.
Open the Enterprise app, start a job, and drag the export ZIPs onto the dropzone — whole, unmodified. Nested ZIPs are expanded automatically, junk entries are skipped, and the full folder paths from your export are preserved as document names. Enter the data subject's details, and detection runs with the subject's information preserved and third-party personal data flagged for redaction.
Two details that matter at DSAR scale: the item report spreadsheets in the export are processed too — they contain names and subject lines, and manual processes almost always miss them. And exports are deduplicated by content: on a monthly cadence, drop the entire new export and only new documents are processed. Previously handled files are skipped automatically.
Bring one real export. We'll process it together, walk the review, and hand you the redacted package and log.
Talk to us Read the full DSAR guide