The ProSearch Pulse Volume 27

This article opens a five-part series on the planning decisions legal teams should make before the Relativity Server transition deadline. The series will walk through five questions that should guide readiness planning, starting with the most immediate one: what should organizations do now to avoid rushed decisions later?

Relativity has announced that organizations will no longer be able to create new matters or workspaces in Relativity Server as of December 31st, 2027. For legal teams, that means the time to move from awareness to active planning and execution is now.

The organizations best positioned for the end of 2027 will be the ones that use the time between now and then to assess their current practices, reduce avoidable risk, and make transition decisions before timing limits their options.

Effective planning starts with a clear understanding of data collection practices, the Server data footprint, matter lifecycles, workflows, governance requirements, RelativityOne cost implications, and future technology priorities. It also requires coordination across legal, litigation support, information governance, compliance, IT, and, in some cases, procurement.

The EDRM community has released EDRM 2.0, a redesigned version of the Electronic Discovery Reference Model that has anchored eDiscovery practice since 2005. About 150 multidisciplinary practitioners built the update. It reflects modern workflows, governance expectations, and the growing role of analytics and AI.

Key insights

• EDRM 2.0 places the Information Governance Reference Model as a foundational layer beneath the discovery lifecycle. Governance drives downstream discovery decisions.
• Identification, Preservation, Collection, and Processing are consolidated into a single Data Acquisition framework. Upstream work is increasingly integrated, especially for producing parties.
• Disposition becomes a discrete core phase. End-of-lifecycle handling becomes a practical, legal, and ethical responsibility.
• Continuous Analysis spans every phase. Legal judgment, plus analytics, data science, and AI, operate throughout discovery rather than at a single step.

For legal teams and vendors, EDRM 2.0 points toward more defensible information governance, earlier alignment on acquisition strategy, and iterative analysis across the case timeline. EDRM also aims to accelerate adoption by licensing the diagram under CC BY 4.0 and publishing public-comment responses for transparency. More details and trustee responses are available here: https://edrm.net/2026/08/edrm-2-0-public-comments-key-issues-and-trustee-responses/

EDRM 2.0 now treats disposition as its own phase: a systematic, defensible process to retain, delete, transfer, or return data after use. That is the cleanup work Kenneth Spencer puts at the start of Relativity Server planning in this issue. He says teams should use 2026 to assess current practices and the Server data footprint before the 2027 cutoff on new matters limits their options.

Relativity is pulling more legal operations work into Microsoft 365 Copilot. Relativity says RelativityOne now integrates with Copilot through the Model Context Protocol (MCP), so teams can run administrative workflows and generate reporting with natural language while governance controls stay inside RelativityOne.

What is driving this shift

Copilot adoption is becoming a default in many firms. ILTA’s 2025 Legal Technology Survey found Microsoft Copilot is used by 68% of firms and 84% of the largest global law firms. That adoption curve creates pressure to meet legal teams where they already work, inside Microsoft 365.

What the MCP integration enables

Relativity’s MCP integration cuts context switching by letting users take RelativityOne administrative actions from Copilot.

• Natural language prompts can trigger administrative workflows
• Reporting can be generated without leaving the Copilot interface
• Governance stays anchored in RelativityOne

Feature coverage by Copilot surface

Functionality varies by Copilot surface.

• Copilot Cowork supports orchestration and reporting
• Copilot Chat supports reporting today
• Copilot Chat orchestration is planned for later this year

How Relativity is positioning Copilot vs. aiR

Relativity is separating the two layers of capability.

• Copilot serves as the plain-language layer for administration
• Relativity aiR handles higher-stakes legal work that needs deeper matter context

What this means for legal teams

For legal ops, the value is speed and standardization across the admin plane.

• Faster setup and management of workspaces
• Better visibility into access and usage patterns
• Reporting that supports governance across matters, clients, and practice groups

The next proof point is how far Copilot Chat orchestration expands when it arrives later this year, and what controls firms can apply in regulated scenarios. From what we see across eDiscovery teams, these integrations work best when early planning sets clear Smart Review Boundaries, aligns stakeholders as One Team, and keeps AI + HI decision rights explicit so administrative convenience does not outpace governance.

Chris Brown, Relativity's president, said Copilot is one more place users can take action in RelativityOne in plain language without leaving Microsoft 365. That capability is RelativityOne-only, and it is the kind of future-technology decision Kenneth Spencer says should shape a Server transition in 2026, while teams can still choose, rather than wait until December 31, 2027, when new matters on Server stop.

California lawmakers have approved SB 574, a measure Reuters calls a first-of-its-kind state law defining how lawyers can use generative AI. The bill now goes to Governor Gavin Newsom for a sign-or-veto decision, as courts deal with a growing number of filings containing unverified AI errors and fabricated citations.

What SB 574 would require

• Lawyers would be prohibited from delegating the practice of law to generative AI and would need to take “reasonable steps” to verify AI-generated material, including case citations.
• Lawyers would be required to correct any false or “hallucinated” output that appears in work product they use, tracking sanctions judges have issued in recent AI-related incidents.
• A transparency provision would require disclosure of generative AI use for documents submitted in court.
• Confidentiality rules would tighten. Lawyers would be barred from entering “confidential or other nonpublic information” into certain generative AI systems, and arbitrators would be prohibited from delegating decision-making to AI.

Why it matters

Supporters, including bill author Tom Umberg, argue that courts rely on the integrity of advocates and that pointing to AI “hallucinations” does not excuse deficient filings. Some experts see substantial overlap with existing professional duties, with Wayne Stacy of the University of California, Berkeley School of Law calling the proposal largely duplicative of current ethics rules. (Reuters)