Enterprise Demo — SafeRedact DSAR Redaction

The Scenario

ACME Corp has received a UK Data Subject Access Request from David Mitchell, a software engineer. The response documents contain his personal data mixed with third-party PII from colleagues, clients, and contacts.

David's data must be preserved. Everyone else's must be redacted.

Data Subject

David Mitchell

✉ david.mitchell@acmecorp.co.uk copy

✉ david.mitchell.personal@gmail.com copy

☎ 07442 839 201 copy

☎ +44 20 7946 0821 (work) copy

# QQ 12 34 56 B copy

# Passport: 549278431 copy

# EMP-2847 copy

◉ 14/03/1988 copy

⌂ 42 Kensington Gardens, Flat 3B, London SW7 4QR copy

Download Demo Pack

20 files · 24 KB · 100% synthetic data

What's in the pack

File	Type	Content	PII
Emails — MSG (Purview eDiscovery format)
msg-001-performance-review.msg	MSG	Annual review — salary, NI number, emergency contact, Revolut offer	8
msg-002-client-breach.msg	MSG	Client breach — bank accounts, sort codes, addresses, DOBs	14
msg-003-restructuring.msg	MSG	Redundancies, severance, NI numbers, promotions	12
msg-004-onboarding.msg	MSG	New starter docs — passport, bank details, tax code	10
msg-005-expense-claim.msg	MSG	Expense approval — client dinner attendees, hotel booking	6
msg-006-disciplinary.msg	MSG	HR follow-up — informal complaint, colleague contact details	4
msg-007-payslip-query.msg	MSG	Payslip breakdown — tax, NI, pension, bank details	7
msg-008-project-handover.msg	MSG	Handover notes — client contacts, credentials, contract value	8
msg-009-reference-request.msg	MSG	External reference request — salary confirmation, confidential	5
msg-010-insurance-claim.msg	MSG	Health insurance — dependant DOB, GP details, treatment info	9
msg-011-it-security.msg	MSG	Security incident — data handling breach, client contract refs	6
msg-012-leaving-party.msg	MSG	Social email — colleague bank details for collection, personal chat	4
Teams Conversations — HTML (Purview export)
teams-chat-engineering-standup.html	HTML	Daily standup — names, client refs, personal email, phone numbers	10
teams-chat-project-neptune.html	HTML	Project chat — database credentials, budget figures, shelving decision	6
teams-chat-hr-private.html	HTML	1:1 with HR — competing offer, counter-offer, salary negotiation	8
Teams Meeting Transcripts — TXT
teams-meeting-transcript-security-review.txt	TXT	Security review — VPN provider contacts, admin access issues, customer data in test DB	12
teams-meeting-transcript-sprint-planning.txt	TXT	Sprint planning — client contacts, medical appointment details	8
Attachments — CSV / TXT
attachment-employee-directory.csv	CSV	HR export — 8 employees with NI numbers, DOBs, addresses, salaries	48
attachment-board-minutes-2026-02-21.txt	TXT	Leadership minutes — restructuring, client contracts, retention counter-offer	15
attachment-incident-log-SEC-2026-0042.txt	TXT	Security incident — data handling breach, remediation actions, vendor contacts	10

All data is 100% synthetic. No real individuals are represented.

How to run the demo

What’s in the pack

Every file is synthetic. ACME Corp is fictional, the National Insurance numbers use the reserved QQ prefix that is never issued to a real person, and the telephone numbers sit in Ofcom’s reserved ranges. Open any of them before you run anything — you should not have to take our word for it.

PDF acme-offer-letter-david-mitchell.pdf Offer letter — NI number, date of birth, salary, home address 3.1 KB
DOCX acme-grievance-statement-GRV-2026-0114.docx Grievance statement — special-category health data 9.6 KB
XLSX acme-payroll-extract-q1-2026.xlsx Payroll extract — sort codes and account numbers 6.5 KB
CSV attachment-employee-directory.csv Employee directory — the subject alongside seven colleagues 1.7 KB
CSV attachment-client-contacts.csv Client contacts — third-party PII only 872 B
CSV attachment-expense-report-q4.csv Expense report — almost no PII; tests over-redaction 982 B
EML email-001-hr-review.eml HR review email (.eml) 1.4 KB
MSG email-001-hr-review.msg The same email as Outlook .msg — proves parser parity 6.0 KB
MSG msg-006-disciplinary.msg Disciplinary record 5.5 KB
MSG msg-009-gdpr-request.msg A GDPR request, inside the export 6.5 KB
HTML teams-001-standup.html Teams chat — engineering standup 8.6 KB
HTML teams-chat-hr-private.html Teams chat — private HR channel 7.5 KB
TXT attachment-access-log-extract.txt System access log — IP addresses and timestamps 2.1 KB
TXT attachment-incident-log-SEC-2026-0042.txt Security incident log 1.7 KB
TXT attachment-board-minutes-2026-02-21.txt Board minutes — prose, sparse PII 1.8 KB

One thing worth knowing. The sample pack runs without a Data Processing Agreement because it describes no real person. If you edit these files or add your own, they stop being sample data — SafeRedact will ask for a signed DPA before processing them. That is the same protection that applies to your live DSAR exports.

Download the demo pack

Optional. The app can load these for you — but download them if you want to inspect the data before it is processed, or keep a copy. 15 documents: PDF, Word, Excel, Outlook messages, Teams chats, spreadsheets and logs.

Open the enterprise app

Go to saferedact.app/enterprise/app and sign in with your pilot account.

Set DSAR mode with subject details

Select DSAR — Preserve Data Subject. Enter: David Mitchell, david.mitchell@acmecorp.co.uk, NI: QQ 12 34 56 B. Use the copy buttons in the scenario card above.

Upload the sample files

Fastest path: open the app and press Run the guided demo — it fills the data subject and loads the pack for you. Prefer to do it by hand? Drag the ZIP onto the upload area, or drop individual files. The app accepts PDF, DOCX, XLSX, MSG, EML, HTML, CSV, TXT and ZIP — matching Purview eDiscovery export formats.

Review AI detections

AI identifies third-party PII while preserving David Mitchell's personal data. Review, approve, or reject each detection.

Export redacted files

Download permanently redacted output. Third-party PII is destroyed via pixel-burn. David's data remains visible.

What to look for

✓ Preserved (data subject)

David Mitchell's name, email, NI number, DOB, address, employee ID, salary — all should remain visible in the output. This is his data and he has a right to see it.

✗ Redacted (third parties)

Marcus Bell, Linda Thompson, Sophie Chen, Kieran Walsh, Priya Sharma, Rachel Okonkwo, James Worthington, Aisha Patel, Thomas O'Brien — all names, emails, phones, NI numbers, addresses, and bank details should be flagged for redaction.

⚑ Confidential (business data)

Project Neptune budget (£340K), restructuring decisions, severance amounts, competing offer from Revolut. Standard DSAR mode focuses on personal data — these items demonstrate where custom category detection adds value.

Architecture

Browser-based

Documents are processed in your browser. Files are never uploaded to SafeRedact servers.

Zero retention

AI analysis uses Anthropic's API under commercial terms: content is not used for model training and is deleted within 30 days. SafeRedact itself stores nothing.

Pixel-burn

Redacted content is physically absent from output files. Not hidden — destroyed. Irreversible.

Read the full security whitepaper →

Ready to evaluate?

Request a pilot account to process your own documents.

Request Pilot Account Launch App →

Try DSAR redactionwith sample data

What’s in the pack

Ready to evaluate?

Try DSAR redaction
with sample data