Personal/Privileged Info and Redactions | Webinar Recording

eDiscovery at Lunch is our monthly drop-in series designed for legal and review teams who want to sharpen their Nuix skills and knowledge, build confidence with legal content, and better collaborate with their team.

Personal/Privileged Info and Redactions

Summary

In this session, Duncan Fraser from Noticia delivered a practitioner-focused deep-dive into redactions in Nuix Discover, covering both civil litigation and Freedom of Information (FOI/ATIP) contexts. The session assumed basic familiarity with Nuix and focused on how teams can identify, apply, track, and produce redacted documents efficiently and defensibly.

Attendees learned how to structure a redaction workflow from start to finish — including case setup, manual review, search term families, AI-assisted PII identification using Noticia Go, quality control, and final production of burned-in redacted documents.

Key Areas Covered

Legal Context for Redactions:

• In FOI/ATIP, the obligation is to produce as much as possible while severing exempt information (personal info, trade secrets, national security, etc.).

• In civil litigation, redactions are largely common-law based — covering irrelevant sensitive content, legal privilege, commercially sensitive information, and sensitive personal information not relevant to the case.

Planning and Case Setup:

• Before starting, define what you are redacting, at what granularity (word, sentence, paragraph), and how redactions will be labelled per exemption type.

• Configure annotations in Case Setup. Coloured highlights (rather than plain black redactions) are recommended — colours are visually easier to track during review and QC.

• Add workflow fields for New Redaction Terms, AI Redaction Terms, Redaction Notes, and QC/Redaction Status to support tracking and sign-off.

Document Preparation:

• Limit the review set to documents that actually require redaction treatment. Documents must be imaged (PDF) before highlights can be applied — use Tools > Imaging for native files.

• OCR is required for text-search-based redacting. Documents without extracted text can still be manually redacted (drawn boxes) but cannot be searched.

• Audio/video and Excel spreadsheets require separate handling — dedicated sessions are available on request.

Find and Redact & Search Term Families:

• The Find and Redact sidebar (images only) supports word/phrase, regex, search term family, and AI entity type modes. Search term families are recommended for any set larger than a handful of documents.

• Use the Test Term feature to catch spelling variants before finalising your list. Terms can be imported from Excel — useful for recycling terms captured during manual review.

AI-Assisted Redaction with Noticia Go:

• Run a bulk scan to identify PII categories (e.g., politicians, vulnerable individuals) and write results to the AI Redaction Terms field. Download, parse, and import into a search term family for bulk application.

• AI output also produces a reportable list of PII found across the set — useful for client validation and documentation of what was identified and protected.

QC and Production:

• Run a final QC pass using search term family highlights to catch missed instances. Track who applied redactions and who signed off on QC for defensibility.

• Produce via custom export as an endorsable image (PDF). Burn in annotations on the export — specify colour (white with black border, or solid black) and label per annotation type. No native files or extracted text should be included in the redacted output.