Our Approach: A Five-Step Process


BeyondRedaction’s five-step approach to redaction enables clients to achieve appropriate, accurate and accelerated sensitive data redaction for FOIA, privacy and privilege requirements.

Leveraging advanced visual classification technology to evaluate sensitive data and redact based on either word or pattern matching or on redacting certain zones within groupings of visually-similar documents, BeyondRedaction allow corporations to quickly and cost effectively redact sensitive data at a rate far exceeding traditional standard redaction technologies and processes.

Step 1: Ingest

Receive and load electronically stored information (ESI) to be considered for sensitive data evaluation and redaction.


During this step, BeyondRedaction technicians help clients prepare and share the ESI requiring sensitive data evaluation and redaction.  The ingestion process consists simply of:


  • Reception of ESI from client via FTP or any secure digital storage device.

  • Loading of ESI by a BeyondRedaction technician into the BeyondRecognition processing platform.


All ESI is received and handled in a forensically sound manner and tracked throughout the entire evaluation and redaction process to ensure defensibility of process and results.

Step 2: Process

Process and create four types of files that enable the appropriateness, accuracy and accelerated results of the BeyondRedaction approach.


During this step, client provided ESI is processed with BeyondRecognition’s visual classification technology with the goal of creating four specific file types for each ESI file.  These four file types include:


  1. Positional Files

  2. Image over Text Files

  3. Text Files

  4. TIFF Files


The creation of these four file types for each ESI file provide all of the inputs necessary to maximize the power and precision of BeyondRedaction technology.

Step 3: Prepare

Development of a single column CSV file that contains sensitive data to be identified and redacted.


During this step, BeyondRedaction technicians work with clients to prepare a single CSV file that contains a listing of sensitive data to be identified and redacted. This listing contains redaction guidance in one of three query forms:


  1. Text (Example: Specific Word or Text Element)

  2. Patterns (Example: Social Security Number)

  3. Strings (Example: Use of Words or Combination of Words)


These three query forms, represented in a single CSV file, serve to guide the actual redaction step of the data-driven redaction protocol.

Step 4: Redact

Processed files (Step 2) and CSV listing (Step 4) are used by the BeyondRedaction platform to rapidly and accurately redact sensitive data.

Example Redaction Based on Text Coordinates

1TextCoordinates (1)

BeyondRedaction works by reading the word dictionary created by BeyondRecognition’s Global Glyph Catalog and applying positional logic to seek out each user-designated word or pattern as it appears on every page of every document in the review set. Redactions can be specified on a per-word basis or based on a pattern, such as social security and credit card numbers, using custom regular expressions. Further, a redaction log is automatically created logging the reason for the redaction, if specified, which can be saved into a format compatible with Microsoft Excel.

During the redaction step, the BeyondRedaction platform initiates redactions at a rate that can scale to over 700,000 redactions per hour. Redactions take place in the following manner:


  1. TIFF Files Redacted

  2. Edited Text Files Redacted (Takes Words Out of File, No Need to re-OCR File)

  3. Image Layer of PDFs Redacted

  4. Text Layer of PDFs Edited


Upon completion of these redactions, the BeyondRecognition platform generates a redaction log that includes:


  1. Name of the File Redacted

  2. The Page/Pages Redacted

  3. The Number of Redactions Per Page

  4. The Reason for the Redaction (Privilege, Privacy)


Redactions can be accomplished as part of a company’s ongoing information governance processing and classification efforts or can be provided as a standalone service to support privacy and legal requirements.

Step 5: Produce

Produce for clients four deliverables for use in addressing FOIA, privilege or privacy requirements or for use in further legal discovery processes.

Example Redaction Based on Page Coordinates

1PageCoordinates (1)

During this step, the BeyondRedaction platform creates for clients a set of documents containing:


  1. Redacted TIFF Files 

  2. Edited Text Files

  3. Redacted and Edited PDF Files (One Object) 

  4. Redaction Log


Upon completion of this step, the redacted information can be used to support ongoing information governance requirements or used to support follow-on privacy and legal requirements.

The “Differentiator” of BeyondRedaction.

Very few document collections have documents with 100% accurate text. BeyondRedaction’s ability to analyze and group documents based on visual appearance provides far greater assurance that all the terms that need to be redacted are in fact redacted.

Contact us today to learn more.

BeyondRedaction is an advanced redaction services company that leverages visual classification technology from BeyondRecognition to deliver fully automated redaction capability for Freedom of Information Act (FOIA), privacy and privilege requirements.

The BeyondRecognition Network