Question 1

What is WorkLens?

Accepted Answer

WorkLens is a local-first privacy-safe data engine. It does two things from one tool: generate realistic test data without exposing real customers (synthetic data), and find/redact personal information in real files (redaction). Everything runs in your browser — your data never leaves your device.

Question 2

What file formats does WorkLens support?

Accepted Answer

CSV, XLSX (Excel), JSON, and XML. PDF support is coming soon.

Question 3

How does the detection work?

Accepted Answer

WorkLens combines multiple detection methods. Pattern matching finds structured data like emails, phone numbers, and credit cards. Mathematical validation confirms ID numbers (currently Australian TFN, Medicare, ABN — more regions coming). On-device AI catches person names, organisations, and addresses in any language. Smart column analysis recognises common data headings. All methods run together and only the strongest result is kept.

Question 4

What types of personal information can WorkLens detect?

Accepted Answer

Over 20 types including: email, phone numbers (any country), credit card (Luhn validated), person names, organisations, addresses, IP addresses, passport numbers, IBAN, date of birth, and more. Australian IDs (TFN, Medicare, ABN, BSB) have deep mathematical validation. US, UK, EU regions coming soon.

Question 5

Which countries does WorkLens support?

Accepted Answer

WorkLens detects universal PII types (names, emails, phone numbers, credit cards, IP addresses) from any country. Australia has the deepest support with mathematical validation for TFN, Medicare, ABN, and BSB. US, UK, EU, NZ, and Canada regions are coming soon.

Question 6

Why does AI take a moment to load?

Accepted Answer

The AI model is 175MB and runs entirely on your device — not on a server. On your first visit, it downloads once and is saved in your browser. Every visit after that, it loads from cache in 1-2 seconds. Pattern detection (emails, phone numbers, IDs) works instantly while the AI model loads. You get results immediately — AI names appear moments later.

Question 7

Does WorkLens work offline?

Accepted Answer

Yes. After your first visit, everything is cached in your browser. Disconnect your internet and scan a file — it works identically. If you have never visited before and are offline, pattern detection still works but AI name detection needs a one-time download. Once downloaded, it works offline forever.

Question 8

What happens if the AI model fails to load?

Accepted Answer

Pattern detection always works — it catches emails, phone numbers, credit cards, tax IDs, and 20+ structured types using mathematical validation. AI is a bonus layer that adds person name detection. If it fails (slow connection, offline first visit), your scan still produces results. When your connection improves, AI automatically retries.

Question 9

Is synthetic data the same as anonymising real data?

Accepted Answer

No — and the difference matters for compliance. Anonymising real data (also called pseudonymisation or masking) starts with a real customer and transforms their values. Under GDPR and the Privacy Act, the result is still "personal data" because it can in theory be linked back. Synthetic data is generated from statistical patterns — no real customer is the source of any synthetic row. That generally places it outside the legal definition of personal data, which is why regulators often favour it for non-production environments. (Treatment varies by jurisdiction and how the data is generated — you remain responsible for your own assessment.)

Question 10

Can I reproduce production bugs using synthetic data?

Accepted Answer

For most bugs, yes. WorkLens preserves your real distribution shape — numeric histograms, categorical frequencies, foreign-key relationships, and edge cases. Bugs that depend on data shape, joins, or rare-value combinations reproduce reliably. For bugs that depend on specific real values (e.g. a customer name with an unusual character), switch to "morph mode" inside Advanced — same engine, deterministic per-entity replacement that keeps real edge cases while removing real identities.

Question 11

How is WorkLens different from Mostly AI, Tonic, or Gretel?

Accepted Answer

Two main differences. (1) Local-first: WorkLens runs in your browser, so your schema, your column names, and your real rows never leave your device. Mostly AI, Tonic, and Gretel are cloud-based — you upload first. (2) Multi-product: WorkLens combines synthetic data, PII redaction, and deterministic morphing in one engine, so you procure one vendor instead of three. Mostly AI is the open-source SDK ceiling if you can run Python; WorkLens is the no-install zero-upload choice for compliance and business teams.

Question 12

What does APRA, GDPR, or the Privacy Act consider as "synthetic"?

Accepted Answer

All three regulators distinguish between pseudonymised and anonymised data. APRA CPS 234 expects you to control non-production data — synthetic data (where no real customer is the source) is the cleanest evidence. GDPR Recital 26 explicitly excludes anonymous information from its scope. The Australian Privacy Act follows similar logic. WorkLens generates true synthetic data, plus a signed audit receipt that proves on-device generation — so you can show the auditor, not just claim it.

Question 13

How is morphing different from synthetic data?

Accepted Answer

Synthetic data is generated from scratch — no real customer is the source of any output row. Morphing keeps your real data and swaps each value for a fake one (e.g. "Alice Wong" → "Marcus Chen") deterministically: the same input always produces the same output, so joins and FKs still work across files. Synthesise when you need the legal cleanness of "no real customer ever existed." Morph when you need exact production fidelity (bug reproduction, FK-dependent analytics, vendor sharing). WorkLens offers both — morph lives inside Synthesize as an Advanced output mode.

Question 14

Does my data leave my device?

Accepted Answer

No. Your file is processed entirely inside your browser. There is no server, no upload, no cloud. You can prove it: disconnect your internet and scan a file — it still works.

Question 15

How can I be sure no data is sent anywhere?

Accepted Answer

Two easy ways to verify: (1) Turn off your Wi-Fi and scan a file — everything works identically. (2) Open your browser's Network tab (F12) during a scan — you will see zero requests containing your file data.

Question 16

What happens to my file after I close the tab?

Accepted Answer

It is securely erased from your browser's memory. No trace of your file remains in any storage on your device.

Question 17

Is WorkLens open source?

Accepted Answer

The detection engine is built using open source technologies. The application itself is proprietary. See our Licenses page for details.

Question 18

What does the AI detect?

Accepted Answer

The AI finds person names, organisation names, and addresses — things that pattern matching alone cannot catch. It runs entirely on your device. No data is sent to any external service.

Question 19

Why is AI optional?

Accepted Answer

The AI needs a one-time download (about 15 seconds on a typical connection). Pattern matching works instantly without it. We give you the choice: quick scan for structured data like IDs and emails, or a full scan that also catches names and addresses.

Question 20

How accurate is the detection?

Accepted Answer

For ID numbers (credit cards, tax IDs, government numbers), accuracy exceeds 97% — these are mathematically validated, not guessed. For names and addresses, AI detection is 80-88% accurate. Every finding shows a confidence level so you know what to review first.

Question 21

Can the AI detect names in ALL CAPS or non-English names?

Accepted Answer

Yes. The AI detects full names regardless of formatting — "JOHN SMITH", "john smith", or "Dr. Priya Sharma" are all recognised as person names.

Question 22

What do the confidence levels mean?

Accepted Answer

Critical: mathematically verified identifiers like credit cards, tax IDs, and government numbers. Sensitive: names, phone numbers, and addresses confirmed by AI or multiple checks. Possibly Sensitive: items that may be personal information depending on context — worth a quick review.

Question 23

How does WorkLens avoid false alarms?

Accepted Answer

ID numbers are validated using official mathematical checks — a random 9-digit number won't be flagged as a TFN. Numeric columns like prices and quantities are automatically skipped. When multiple detection methods agree on the same item, confidence goes up. The result is a false alarm rate under 5%.

Question 24

What does "Scan Incomplete" mean?

Accepted Answer

It means AI has not been enabled, so only pattern-based detection was used. Names, organisations, and addresses are not checked without AI. Click "Enable AI Scan" in the top bar for a complete scan.

Question 25

Can I manually mark or unmark cells as PII?

Accepted Answer

Yes. Click any cell to mark it as PII with a severity level of your choice. Click a detected cell to unmark it. You can also redact entire columns or all matching values across the file.

Question 26

What does the redacted file look like?

Accepted Answer

Personal information is replaced with labels — for example, "John Smith" becomes "[Person Name]" and "123 456 789" becomes "[Tax File Number]". The rest of your file stays exactly the same.

Question 27

What is the audit receipt?

Accepted Answer

Every scan produces a signed receipt that contains two things: a fingerprint of your file and detections, and a digital signature that proves WorkLens generated it. Think of it like a notarised document — the fingerprint proves what was scanned, and the signature proves who signed it. Download it as PDF or JSON from the right sidebar after any scan.

Question 28

What is the difference between the hash and the signature?

Accepted Answer

The hash locks the content — it proves this exact file with these exact detections was scanned. Change one character and the hash is completely different. The signature locks the signer — it proves this receipt was generated by WorkLens on a specific device. You need both: the hash alone can be recomputed by anyone, but the signature cannot be forged without the private key stored on the original device.

Question 29

How do I verify a receipt?

Accepted Answer

Go to theworklens.com/verify and drop your receipt JSON file. WorkLens checks the digital signature against the hash — if they match, the receipt is authentic and nothing was tampered with. You can also drop the original file to verify it has not been modified. Everything runs in your browser — the receipt is never uploaded.

Question 30

Can someone fake an audit receipt?

Accepted Answer

No. To forge a receipt you would need the private signing key, which is locked inside the original device's browser storage. Without it, creating a matching signature is mathematically impossible — it would take longer than the age of the universe to brute-force. If anyone modifies the receipt after it was generated, the signature verification will fail.

Question 31

Where is the signing key stored?

Accepted Answer

The private signing key is generated on your device and stored in your browser's secure storage. It never leaves your machine. When signing, the key is loaded as non-extractable — meaning even scripts running on the page cannot read the raw key. Each device has its own unique key.

Question 32

How large can my file be?

Accepted Answer

CSV files work best under 2 MB (up to 10,000 rows). XLSX under 3 MB (up to 5,000 rows across multiple sheets). JSON and XML under 1 MB (up to 5,000 properties/elements). Maximum file size is 10 MB. A desktop version with higher limits is planned.

Question 33

Is WorkLens free?

Accepted Answer

Yes. The scanner is completely free with no signup required. Premium features for teams and enterprise compliance are coming soon.

Question 34

Will there be an enterprise version?

Accepted Answer

Yes. Enterprise plans will include compliance policy packs for regulated industries (finance, healthcare), team management, and options for on-premise deployment.

Question 35

Do I need to create an account?

Accepted Answer

No. Just drop a file and scan. No signup, no login, no account required.

Question 36

What browsers are supported?

Accepted Answer

Chrome, Edge, and Brave work fully. Firefox and Safari have partial support.

Question 37

Does WorkLens work offline?

Accepted Answer

Yes. After your first visit, everything is saved in your browser. Disconnect your internet and scan — it works identically.

Question 38

How is this different from cloud-based tools?

Accepted Answer

Cloud tools upload your file to a server for processing. WorkLens processes everything inside your browser — your file never leaves your device. There is no server to hack, no data to leak, and no vendor to trust.

Frequently Asked Questions