Job personal-details-masked-in-a-spreadsheet-with-recall-measured · revised 12 October 2026
Mask names, emails and phone numbers in my spreadsheet, and measure what was missed
A script you run swaps names, emails and phone numbers for consistent tokens. Misses are measured on a set we plant and made-up details you add. Names are tried, not pass-marked. No guarantee.
You might be seeing
- Free-text columns hold names, emails and phone numbers mixed with useful comments
- Nobody can say how many personal details a manual clean-up misses
No passwords, keys, card details or admin invites needed to start.
What usually happened
Masking personal details is useful and cannot be guaranteed: even a widely used open tool says there is no guarantee it finds all sensitive information. A 'redacted' file that still holds a name is worse than a labelled risk. The honest version measures what was missed and says so.
Who it’s for: Analyst or office manager who wants to use an AI tool on a spreadsheet of customer or staff comments but must remove personal details first.
Usually starts when: A spreadsheet of free text is ready for analysis, and nobody is comfortable that the names and contact details are gone.
The result: A script you run replaces names, emails, phone numbers and ID-like numbers with consistent tokens, so the same email or phone number gets the same token, and the key stays with you. Misses are measured on a planted synthetic set we make (a smoke test) and on made-up details you add to your own copy, and a sample of your own rows is checked by your person. It is not a guarantee.
Check whether this job fits
Answer these without sending files or logins. Nothing is submitted unless you choose to contact us.
What you get
- The masking script and run steps
- The token key file, created on your side
- Misses by type on the held-back planted set
- Your 100-row sample check, recorded
- A plain-words page on what the script cannot catch
Included
- A script for one operating system, run by you on your own computer
- One spreadsheet of up to 50,000 rows with up to 10 text columns
- Pattern types: names, emails, phone numbers and ID-like numbers
- A planted synthetic set of 300 rows, 150 held back with at least 100 planted items of each of emails, phone numbers and ID-like numbers, to measure misses
- A sample check of 100 of your own masked rows by your named person
Not included
- A guarantee that all personal details are removed
- Reversing tokens for anyone but you
- Advice on whether you may process the data
- Images, attachments or audio
- Sending your spreadsheet anywhere
How we know it’s done
Agreed with you before work starts. Each check produces evidence you keep.
On the held-back planted set, at least 95 of 100 planted items of each of emails, phone numbers and ID-like numbers are masked, reported by type with the list of misses (95 of 100 shows the true rate is at least 89% with 95% confidence). We plant this set ourselves, so it is a smoke test of the script. In a copy of your own spreadsheet you add 20 made-up emails and 20 made-up phone numbers that we are not told about; at least 18 of each are masked (at least 71% with 95% confidence). Names are reported but not passed or failed.
Evidence: The misses table with the list of misses.
The same planted email or phone number always gets the same token, and different ones get different tokens.
Evidence: A consistency check on 20 planted repeats.
Run with the network blocked, the script gives the same output, and a connection monitor shows no outbound connection.
Evidence: Both outputs and the monitor log.
Your named person checks 100 of your own masked rows, and the number of personal details found is recorded. No pass mark is promised for this check.
Evidence: Your recorded check.
The token key file is not in the delivered script or output files.
Evidence: A search of the delivered files for the key.
Sign-off. You inspect the misses table, the no-network test and your own sample check, and sign off in writing. Payment follows sign-off. Whether the masked file is fit to use is your decision.
If it fails. If the held-back test is not passed, you do not pay for this fixed scope. You keep the results, the failures and the test set, and we agree whether to stop or re-quote. No surprise work.
When it fits, and when we stop
It fits when
- You can run a script on your own computer, or a colleague can
- The spreadsheet has up to 50,000 rows and 10 text columns
- A named person checks 100 masked rows and decides whether the result is fit to use
- You accept that a clean report is not proof that no personal details remain
We stop and tell you if
- You want a guarantee or a compliance certificate
- You want your spreadsheet sent to us to be masked
- The text is mostly in languages other than English
- Special-category details, such as health, appear throughout the text
What could go wrong
We receive no files from you. The script reads your spreadsheet on your computer and writes a masked copy; the original is not changed.
Scroll the table sideways to read it all.
| Risk | How we handle it |
|---|---|
| A masked file is treated as safe and still holds a name. | The report states the misses by type, the limits page says so, and your person checks 100 rows before use. |
| Tokens can be reversed by others. | The key file is created on your side and never sent to us. |
| The script sends data out. | A no-network test is part of the acceptance. |
A separate automated reviewer checks the misses table and the token design before hand-over and does not mark the held-back test. A named person on your side runs the script, checks 100 masked rows and signs off. No person on our side is included unless your agreement names one, and at launch the work is largely automated.
How we deliver
We arrange the work and independent review, then show you the result against the agreed checks. You keep authority over your systems.
- Agree the pattern types, the pass marks and the severe errors
- Build a synthetic set of 300 rows with known planted details
- Build the masking on 150 rows; keep 150 held back
- Measure misses by type on the held-back rows; run the no-network test
- Independent review of the misses table and the token-key design
- Hand over the script and steps; you run it and check 100 rows
This is a one-off job, not emergency cover or a subscription. We confirm eligibility, the price, a start window and a delivery date before you accept, and work starts only after inputs, a secure handoff route and permissions are in place. Model usage up to a cap agreed in writing is included, and an enquiry creates no charge or booking.
Need to keep it working?
Want the masking re-tested when your data changes? Ask about a re-check.
Ongoing work is separately scoped and quoted: no monitoring, response-time guarantee or automatic subscription is included in this job.
Explore an ongoing managed lane, or mention the responsibility you need in your enquiry.
What you can check
This is a new service. We have not delivered this job for a client yet.
Other ways to get this done
- Microsoft Presidio is a free open-source toolkit for finding and masking personal data. It needs someone who writes code, and it also gives no guarantee. presidio.dataprivacystack.org
- If you only need aggregate numbers, dropping the text columns altogether is safer than masking them.
Questions
Is the masked file anonymous?
Not necessarily. Masking reduces personal details; it does not prove they are gone. Your person checks a sample, and the decision to use the file is yours.
Do you see my spreadsheet?
No. You run the script yourself.
Can the tokens be reversed?
Only with the key file, which is made on your side and stays there.
Send an enquiry
Send us
- What the spreadsheet holds and how many text columns it has
- Which operating system you would run the script on
- No real records, personal data or credentials in the first enquiry
Later, once you agree
- Nothing from your spreadsheet: we build and test on our own synthetic rows
- A named person who runs the script and checks 100 rows
Your spreadsheet never leaves your computer. We build and test the script on synthetic rows we create, hand it over, and you run it yourself. The token key file is made on your side and stays there.
Email fallback: open your mail app
If website submission is unavailable, review and send the fallback email yourself. An email fallback is not a website receipt. Or write to hello@syntheticindustry.ai with “personal-details-masked-in-a-spreadsheet-with-recall-measured” as the subject.