Reflections from RelativityFest 2023

• October 10, 2023
• Gulfstream Legal

Gulfstream recently sent a team of employees to the 2023 RelativityFest to learn and network with other Relativity experts. Ryan Rhodes, our VP of Business Development, attended several sessions and provided the following summaries upon his return from the conference. If you couldn’t attend, read on for his highlights.

Key Announcements by CEO, Phil Saunders:

• Emphasis on strengthening relationships with channel partners and end-users.
• Plans for both Relativity Server and RelOne were detailed. Relativity Server will continue its development, especially for features not requiring the cloud. RelOne will expand with features in Relativity Contracts, PII, Collect, Legal Hold, and Data Breach Response workflows. Outside of these features, the two versions will remain similar.

Future of Reviewing in Relativity:

• Introduction of a private Large Language Model (LLM) within a firm’s Relativity instance, located in a private cloud.
• Features include automated reviews, specific query responses, and high accuracy in document reviews.
• This innovative approach, termed “Relativity Air,” will begin its rollout next year.
• Air will only be supported in RelOne.

Session #1 – Relativity Contracts 101 Summary

This session dealt with contract review and how Relativity helps corporations better manage their thousands of active contracts.

Regulatory Response Services:

• Build custom models to find, extract, and compare key language in agreements to help clients ensure ongoing compliance and avoid penalties.
• Pricing is based on a one-time-per-document unit model.

Corporate Transaction Services:

• Automatically categorize agreements and their sections by type to help clients understand the obligations, liabilities, and potential risks involved in M&A activity.
• Pricing levels are aligned to Flex Commit tiers.

Contract Data Management Services:

• Use purpose-built AI to identify and extract datapoint-level attributes from client contracts into a database for monitoring, reporting, and identifying inconsistencies.
• Pricing is based on unique, successfully processed documents run through the product in 25 MB units.

Download the presentation here.

Session #2 – RelativityOne Legal Hold & Collect Summary

This session taught attendees how to preserve, hold, protect and collect data using RelativityOne’s Legal Hold & Collect.

Relativity One’s Flexibility:

• The platform can handle the entire lifecycle of a matter, allowing corporations to choose which processes to handle in-house and which to delegate to partners. The workflow is based on the Electronic Discovery Reference Model (EDRM) and Information Governance Reference Model (IGRM).

Legal Hold / Preservation:

• Emphasized the importance of an effective and defensible workflow for e-discovery.

Why Relativity Legal Hold:

• Integrated into an end-to-end platform.
• Effective custodian communication.
• Flexibility with templates and customization.
• Integrations with O365 & G-Suite.
• Versatility for various use cases, including legal hold, attestations, and task management.

Download the entire presentation here.

Session #3 – Transform Review with Generative AI:

This session covered how Relativity is leveraging Generative AI to transform the review process.

Relativity’s AI Principles:

• Relativity builds AI with a clear purpose that delivers value to customers.
• The AI development emphasizes fairness, privacy, security, and accountability.

Common Myths about Generative AI:

• Addressing misconceptions about Generative AI, including its capabilities, risks of hallucinations, data privacy concerns, and differences between various large language models.

Generative AI:

• Relativity is adopting a cautious approach to understand the value and risks of Generative AI.
• The initial products focus on individual document analysis to build confidence in its capabilities.

Purpose of Building AI:

• The goal is to accelerate the review process, identify key information, and improve privilege review and privilege log creation.

Download the entire presentation here.

Relativity Roadmap and Notes:

Relativity aiR (Review powered by GPT) Highlights:

• The system is transitioning to leverage Natural Language Processing in multiple workflows.
• Matter profiles have been revamped to generate prompts, offer predictions with explanations, and identify crucial documents.
• Enhanced UI now includes prediction markers and highlighted overlays.
• Users can provide feedback, enabling the models to learn and adapt dynamically.
• The foundation is built upon Microsoft Azure OpenAI.
• A conducted study showcased a post-training precision of 85% and a recall rate of 98%.

Additional Noteworthy Changes:

• Workspace Management: Introduced a recycle bin functionality, preserving deleted workspaces for a week and facilitating easy restoration.
• Reporting: New features include cross-matter analytics and insights into billing.
• Legal Holds & Collection: By early 2024, expect a revamped legal hold and collection system for platforms like M365, Slack, and Google. SP collection is anticipated by end of 2023, with direct collection to review workspaces.
• Data Transfer: A modernized automated module is now available, boasting improved efficiency and auto-assignment of custodians.
• Module Enhancements: TextIQ-Heretik modules have been rebranded as Contracts, PI Detect, and Data Breach Response. No updates were shared about potential pricing alterations for these modules. Custom AI modules can be tailored for these tools.
• PI Management: The system can identify, assess, and report Personally Identifiable Information (PI) and has the capability to redact PI automatically prior to finalizing.
• Message Handling: The focus is on efficiently processing and reviewing concise messages. The platform uniquely categorizes messages, setting them apart from standard documents.
• Review Center: The existing Active Learning workflow has been upgraded and enhanced.

Gulfstream’s Vision:

• Partners can opt for either Relativity Server or RelOne without long-term commitments.
• Some clients prefer Relativity Server for slightly better hosting rates, while others (mainly corporations) lean towards RelOne for its extended features.
• Gulfstream ensures expertise in both platforms, offering the latest updates and patches. The payment model remains Pay-As-You-Go (PAYG) without long-term obligations.