Using Noticia Go for AI Review | 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.

Using Noticia Go for AI Review

Summary

In this session, Lucas Fraser and Duncan Fraser from Noticia delivered a practical, practitioner-focused introduction to Noticia Go, our generative AI tool for Nuix Discover. The session focused on how teams can use AI to accelerate document review, perform relevance, privilege, and issue analysis, and handle objective coding more efficiently—all while maintaining control, defensibility, and data security.

Attendees learned how Noticia Go processes individual documents according to user-defined prompts, produces structured outputs that can be written to Nuix fields, and enables bulk scans for large review sets. The team also highlighted practical tips for integrating AI into existing workflows, maintaining quality control, and leveraging AI to reduce repetitive manual work.

Key Areas Covered

Introduction to Generative AI and Noticia Go:

• High-level overview of generative AI (large language models) and how it differs from traditional TAR or other legal AI tools.

• Noticia Go is purpose-built for eDiscovery, reviewing one document at a time for a defined task, and is designed to be fast, consistent, and defensible.

• Emphasis on Canadian data residency and privacy: document content is never stored externally or reused for model training.

Setting Up Noticia Go in Nuix:

• Enabling Noticia Go in a case, connecting output fields, and ensuring documents have extracted text for scanning.

• Optional workspace setup and conditional coding templates to streamline AI review.

• Overview of prompt configuration, history tracking, and quick-start libraries for common tasks.

AI-Assisted Review Workflows:

• Single Scans vs. Bulk Scans: Single scans are free; bulk scans can handle up to 25,000 documents at a time, with processing speeds averaging 200–300 documents per minute.

• Relevance, Privilege, and Issue Review: Prompts can be structured to output to separate fields for maximum flexibility.

• Objective Coding Assistance: Extract dates, titles, people, and document types consistently and accurately.

Live Demonstration Highlights:

• Demonstrated creating dates and summaries for documents missing metadata.

• Showed relevance, privilege, and issue review using a single prompt across a document set.

• Bulk scan execution and field population, including quality control and validation checks.

• Using Noticia Go to deep-dive into long transcripts and generate structured outputs for follow-up questions or analysis.

Defensibility and Quality Control:

• Sampling, consistent application of AI prompts, and manual QC remain critical to ensure defensible review outcomes.

• AI results are repeatable and auditable, with prompt history maintained for transparency.

Practical Value:

• Noticia Go can dramatically reduce time spent on repetitive review tasks, such as coding dates, extracting invoice metadata, or summarizing document content.

• Example: a workflow that might take a human reviewer hundreds of hours can be completed in minutes with AI assistance, while still allowing human oversight.

Follow-Up Note from the Session

One question that came up after the session was whether relevance, privilege, and issue analysis could be run separately. Yes, these can be executed as separate prompts and written to distinct fields—allowing you to choose the approach that best fits your workflow and defensibility strategy.