RelativityOne has rolled out a wide set of enhancements across review, processing, governance, and reporting, with particular momentum in its Artificial Intelligence assisted review suite. In aiR for Review, a new Prompt-Kickstarter for Issues drafts issue descriptions to speed project setup, while Project Sets allow multiple saved searches within a single project to iterate and validate criteria at scale. Teams can now validate prompt criteria directly in Review Center by calculating precision and recall in platform, and a new All Issues tab consolidates outputs for easier assessment. Performance improvements now run aiR for Review and aiR for Privilege jobs about three times faster, job size limits have increased to 150,000 documents per job and 600,000 per instance, and users can export prompt criteria to Word for collaboration. The aiR offerings have also expanded regionally to Japan, India, Singapore, Hong Kong, and South Korea.
On the privilege front, aiR for Privilege is generally available with generative Artificial Intelligence that accelerates first pass privilege review and privilege log creation, providing natural language rationales with predictions. A June update introduced a GPT-backed entity role classification engine that replaces prior steps such as Identify Spammers and Scrub Disclaimers, reducing pipeline complexity while maintaining high accuracy and adding rationales. The product also added Customer-Managed Key support for encryption control. Relativity flagged that the foundational large language model used with its generative Artificial Intelligence solutions was upgraded on June 15, noting that results may differ on jobs run after that change.
Beyond Artificial Intelligence, review and production workflows received notable upgrades. PDF Workflows now support storing outputs in sets with status and error visibility, and users can download stored PDFs with transparent redactions and highlights for quality control. Review gains include Conditional Coding Rules with a refreshed interface, document notes and comments for collaborative review, an A/V Transcription Viewer supporting over 140 languages, and a Short Message Investigative Workflow that introduces an Elasticsearch index, advanced short message search, highlights, and message-level coding. Production can now apply branding directly to stored PDFs. Redact adds a markup validation agent to confirm burn readiness before production and introduces inverse markups to track removed spreadsheet redactions.
Data ingestion and governance also broadened. Collect can now bring in prompts, responses, and metadata from Google Gemini in Relativity Short Message Format, and integrates with Cellebrite Endpoint Inspector and Mobile Now for remote mobile collections. Processing adds support for HEIC files, adopts EDRM Message ID Hash for cross-platform deduplication, and begins an Office 2024 transition that may alter EML appointment hashing; support for Microsoft Publisher files will wind down in 2025. Legal Hold now preserves Slack data in place, exposes error messages, supports token auto-refresh for Graph API, and can retry failed communications. Security Center introduces a security configurations view and updates support group handling, while Data Management ties cold storage strictly to system Workspace State. Reporting and admin improvements include Usage Reports templates and permissions, Review Center queue build visibility and saved models, anonymized reviewer names, background workspace creation, Import/Export field renaming and Artificial Intelligence field mapping, Translate upgrades for stored PDFs, higher job limits, and source language selection, plus accessibility and tooling updates such as a Staging Explorer refresh and a new Data Breach Response project dashboard and entity report.
