Anonymize customer data for AI in administrative work
To anonymize customer data for AI, you start from the documents of an ordinary day in administration, case handling or payroll: a tracking spreadsheet in Excel, a customer file in Word, an invoice in PDF, a ticket export in CSV, payslips. Each one names a customer, an employee, a member, and carries identifiers that point to them unambiguously: customer, order, ticket or payslip number. Biffer recognizes them on your computer, with no connection, replaces them with stable aliases and highlights them in black in the output file.
On this page we go through these files one by one, how to redact them, the GDPR framework and the AI uses that follow.
An administrator's files and what they contain that is sensitive
Administrative and back-office staff receive documents from every department and keep the spreadsheets that link them together. Here are the files we find everywhere, with what each one reveals.
Tracking spreadsheets (Excel, CSV)
One row per customer or order: name, number, amount, status, free-text comments where phone numbers and remarks pile up.
Customer and member files (Word, PDF)
Customer or member number, full contact details, history of the relationship, sometimes tense exchanges.
Purchase orders and invoices (PDF)
Order and invoice numbers, French company number (SIRET), IBAN, signatory, negotiated terms, delivery address.
Complaints and tickets (text, CSV, PDF)
Complaint or ticket number, the requester's name, description of the problem, what was said, response times.
Payslips and employee files (PDF, Excel)
Employee number, payslip number, French social security number, salary, absences, garnishments.
Insurance claims and correspondence (PDF, Word)
Claim number, circumstances, third parties involved, exchanges with the insurer.
Text PDFs, Word, Excel, CSV and plain text go through the same process (see redact a PDF); images and scanned PDFs go through built-in optical character recognition, offline, then are redacted by blacking out the pixels, with no possible restoration.
Redacting these files in practice: formats, aliases, black highlighting
Everything happens on your computer, with no connection (see offline software).
- You drop the file with the "Administrator (admin, files, payroll)" profile enabled. The local analysis highlights people, addresses, emails, dates, SIRET, IBANs, customer, order, ticket and employee numbers.
- You untick false positives (a product name, a workflow status) and add what is missing, all occurrences at once.
- Each item becomes a stable alias (CLIENT_001, COMMANDE_014 for an order, TICKET_002; the alias labels are in French), identical from one file to the next for the same customer.
- In Word and PDF, redacted data is highlighted in black, an option enabled by default: you see at a glance what was removed. In Excel and CSV, columns, formulas and filters are preserved.
- You choose whether to save the mapping table, encrypted (Argon2id and AES-256-GCM) or in plain text. With it, the real names are restored in the AI's answer; without it, the redaction is irreversible.
The queue takes a month of invoices or tickets, in reviewed or automatic mode. Biffer is free for five files per month, all redaction features included, after a one-time email confirmation in the software, with no password. Beyond that, Biffer for Mac or Biffer for PC costs €10 excl. VAT (€12 incl. VAT) per computer, paid once by card via Stripe: a lifetime license for the version purchased, tied to the computer, unlimited files, each major version sold separately. Every feature is included from the free plan (see pricing).
The administrator profile adds customer relationship and payroll identifiers to the general detections. Here is the exact list from the software:
| Field | Where you find it |
|---|---|
| Customer number | Customer files, invoices, tracking spreadsheets, CRM |
| Order number | Purchase orders, confirmations, delivery notes |
| Invoice number | Invoices, reminders, sales journals |
| Complaint number | Letters, complaint registers |
| Ticket number | Support tools, CSV exports, reports |
| Employee number | Personnel files, payroll spreadsheets |
| Member number | Associations, health insurers, group contracts |
| Claim number | Insurance files, exchanges with the insurer |
| Payslip number | Payslips, payroll journals, archives |
On top of these come the fields shared by all professions (file number, reference, contract number, employee ID) and the general detections: people, emails, phone numbers, addresses, SIREN and SIRET, French social security number, IBAN, license plates, dates. An internal code specific to your department can be added by keyword (see anonymization by profession).
GDPR and data minimization: an administrator's obligations
Administrative staff process personal data on behalf of an employer or of clients, who are the controllers under the GDPR and, in France, the Data Protection Act No. 78-17 of 6 January 1978. These texts require a specified purpose, data minimization (Article 5 of the GDPR) and appropriate security measures.
Two consequences for these files. With offline software, the named spreadsheet stays inside the company and you keep control of the question of processors. And identifiers add up: Biffer removes the name, the customer number, the email and the delivery address together, for complete protection. Biffer prepares, you validate: the administrator reviews the preview and has the final word on what is shared.
AI use cases once the files are redacted
A redacted ticket export keeps its reasons, dates and response times: TICKET_002 opened by CLIENT_017 about order COMMANDE_014, resolved in three days. The AI draws statistics from it without knowing anyone.
- Sort a month of complaints by reason and spot recurring causes.
- Draft a reply to a customer from the history of the file.
- Detect duplicates and inconsistencies in a tracking spreadsheet.
- Prepare the monthly activity report: volumes, response times, trends.
The redacted file can be pasted into ChatGPT, Claude or Gemini; the page anonymize a document details the policies per data type. The answer, loaded back into Biffer with the mapping table, gets the real names back, still locally. Biffer is designed in Lyon by Romain Rissoan, its founder, trainer and consultant, at OPCADIA.
Frequently asked questions
How does Biffer handle an Excel spreadsheet with several thousand rows?
Cell by cell, on your computer. In each cell, Biffer spots names, emails, phone numbers, addresses, IBANs and customer, order or ticket numbers, then shows the preview with each detection highlighted by type. You validate by type rather than cell by cell: pseudonymize customers, keep amounts and dates, mask free-text comments. The output file keeps its tabs, formulas, filters and cell formats. The same customer gets the same alias on all their rows, which preserves groupings.
Is an order or ticket number really personal data?
It becomes personal data by association. An order number leads, in your tool, to a customer and their address; a ticket number leads to a requester and what they said. That is why the administrator profile offers these numbers as fields in their own right, with a stable alias (COMMANDE_014, TICKET_002) that keeps the link between rows while protecting the person. If an analysis requires the real number, the "keep" policy applies to that data type alone.
Can you process a CSV export from a CRM or a ticketing tool?
Yes, with no prior conversion. Biffer reads the CSV with its separator and encoding, spots the columns that contain personal data and identifiers, and replaces them with aliases while keeping the exact structure of the file. Free-text columns are analyzed like a document: names, emails and phone numbers mentioned in a ticket are detected too. The output file can be reimported into a spreadsheet or pasted into the AI. A series of monthly exports processed in the same queue shares the same aliases.
Your business documents anonymized, ready to be published or handed to an AI tool.
Biffer runs on your computer without sending a single file. Free for five files a month, unlimited for €10 excl. VAT per computer.
100% offline · Free: 5 files a month · No file sent