Anonymize accounting documents for AI

To anonymize accounting documents for AI, you work on the files the firm produces every day: a balance sheet in PDF, a corporate tax return, an FEC (the French standard audit file of all journal entries) in plain text, a general ledger in Excel, a batch of payslips. Each one names the client, its suppliers, its employees, its banks, and carries precise identifiers: EU VAT number, French company number (SIREN), account and invoice numbers. 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 framework of professional secrecy and the AI uses that follow.

Calculator, rising bar chart and stack of invoices, icon for accountants
Accountants: EU VAT numbers, invoice numbers and client identifiers are detected and replaced.

An accounting firm's files and what they contain that is sensitive

A client file is made of very different formats, from the laid-out PDF to the standardized text file. Here are the ones we find in every firm, with what each one reveals.

Balance sheets and annual accounts (PDF, Word)

Company name, SIREN, registered address, the manager's name, salaries, shareholder current accounts, personal guarantees in the notes.

Corporate tax returns (PDF)

Return number, EU VAT number, identity of the shareholders and share capital breakdown, membership number with an approved management body (OGA).

FEC, general ledger and trial balance (text, CSV, Excel)

Every entry label names a third party: customer, supplier, employee, bank; receivable and payable sub-accounts (411 and 401 in the French chart of accounts) often include the name.

Payslips and DSN payroll filings (PDF, Excel)

Name, French social security number, employee number, salary, the employer's URSSAF number (social contributions), absences and wage garnishments.

Invoices and supporting documents (PDF)

Invoice and voucher numbers, the supplier's IBAN, check numbers, unique SEPA mandate references on direct debits.

Engagement letters and correspondence (Word)

Negotiated fees, the client's situation, exchanges with the tax authorities or the bank.

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.

Download Biffer

Redacting these files in practice: formats, aliases, black highlighting

Everything happens on your computer, with no connection (see offline software).

  1. You drop the file with the accountant profile enabled. The local analysis highlights people, addresses, SIREN and SIRET, IBANs, VAT, invoice and account numbers.
  2. You untick false positives (an expense account mistaken for a voucher number) and add what is missing, all occurrences at once.
  3. Each item becomes a stable alias (PERSONNE_001, SIREN_001, FACTURE_014; the alias labels are in French), identical from one file to the next for the same third party; an FEC stays balanced, only the labels change.
  4. 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 and formulas are preserved.
  5. 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 complete annual file, 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 accountant profile adds accounting and payroll identifiers to the general detections. Here is the exact list from the software:

FieldWhere you find it
EU VAT numberInvoices, tax returns, VAT returns
Account numberGeneral ledger, trial balance, FEC
Invoice number, voucher numberPurchase and sales journals, supporting documents
Tax return numberCorporate tax returns, online filings
URSSAF numberPayslips, DSN filings, letters from the agency
OGA membership numberCertificates, tax returns
Unique SEPA mandate referenceDirect debits, bank statements
Check numberBank journals, deposit slips

On top of these come the fields shared by all professions (file number, reference, contract number, employee number) and the general detections: people, emails, phone numbers, addresses, SIREN and SIRET, French social security number, IBAN, license plates, dates. A client code specific to your firm can be added by keyword (see anonymization by profession).

Download Biffer

Professional secrecy for chartered accountants

In France, the chartered accountant (expert-comptable) is bound by professional secrecy under the Ordinance No. 45-2138 of 19 September 1945 and the code of ethics set by Decree No. 2012-432 of 30 March 2012. Disclosing information covered by this secrecy falls under Article 226-13 of the French Criminal Code.

Two consequences for the firm's files. Secrecy covers the client and its third parties: suppliers, employees and banks that appear in the entries are redacted just the same. Payroll data also falls under the GDPR, since the firm is the controller for the payslips it produces. Biffer prepares, you validate: the professional reviews the preview and has the final word on what is shared.

Download Biffer

AI use cases once the files are redacted

A redacted FEC or general ledger keeps its amounts, dates and structure: the AI sees entries for CLIENT_003 and FOURNISSEUR_012 (supplier), and the real third parties stay protected.

  • Check the consistency of an FEC before year-end closing: duplicates, gaps in numbering.
  • Summarize a balance sheet in plain language for the presentation meeting with the client.
  • Draft the outline of a summary note or an engagement letter.
  • Analyze a batch of invoices to extract payment terms and due dates.

The redacted file can be pasted into ChatGPT, Claude or Gemini; the page anonymize before Claude gives the useful settings. 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.

Download Biffer

Frequently asked questions

How does Biffer handle an FEC or a CSV export from accounting software?

The FEC is a standardized text file, read column by column. Biffer spots third parties in the labels and sub-accounts, voucher and invoice numbers, dates and amounts, then shows you the preview with each detection highlighted by type. You can keep dates and amounts, which any analysis needs, and redact only the labels and identifiers. The output file keeps exactly the same structure, separator and encoding: it stays readable by a spreadsheet or by the AI. The same principle applies to a general ledger exported as CSV or Excel.

Are account numbers really data to redact?

General ledger accounts (512 for the bank, 607 for purchases, 706 for services) are impersonal and can be kept. Sub-accounts are another matter: 411DUPONT or 401MARTIN contain the third party's name in clear, and a chart of accounts is then enough to rebuild the customer or supplier portfolio. That is why the accountant profile offers the account number as a field in its own right, with a stable alias (COMPTE_007) that stays identical in the trial balance, the general ledger and the FEC. You choose the policy per type: pseudonymize sub-accounts, keep general accounts.

Can you get the real names back in the summary returned by the AI?

Yes, if you saved the mapping table at export. It links each alias to its original value and can be encrypted (Argon2id and AES-256-GCM) or in plain text, as you choose. When the AI returns a note that mentions CLIENT_003 or FACTURE_014, you load it back into Biffer: the real values are restored in the final document, on your computer. Without a table, the redaction is irreversible, which suits a single-use file leaving the firm.

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