Sources
1 EDRM, “EDRM Invites Global Community to Shape EDRM 2.0 via Public Comment,” June 30, 2026; Doug Austin, “EDRM 2.0 Model Released by EDRM for Public Comment,” eDiscovery Today, June 30, 2026.
EDRM 2.0: Redrawing the map for discovery in the AI era
A twenty-year-old map meets new terrain
Since 2005, the EDRM, stewarded by George Socha and Tom Gelbmann, has given courts, corporate legal departments, law firms, service providers, and technologists a shared language for how electronic evidence moves from creation to the courtroom. That familiar left-to-right diagram, running from Information Governance through Identification, Preservation, Collection, Processing, Review, Analysis, Production, and Presentation, is arguably the most recognized image in our eDiscovery industry¹.
But the terrain has shifted. Volumes, variety, and velocity of data in the agentic AI era are straining workflows built for a slower, more sequential world. Recognizing this, EDRM has released EDRM 2.0 for public comment. This is the first substantive update since the Information Governance Reference Model (IGRM) was folded in, shaped by roughly 150 practitioners across the global community¹.
What's actually changing
The proposed model preserves the familiar simplicity while reflecting how discovery is really practiced today. Four shifts stand out:
- Information governance becomes the foundation. IGRM moves from a side note to the bedrock layer beneath the entire lifecycle, reinforcing that governance decisions shape every downstream discovery activity.
- The front end consolidates into “Data Acquisition.” Identification, Preservation, Collection, and Processing are grouped together, acknowledging that modern tools increasingly perform these steps simultaneously and in place. A “Volume to Relevance” wave underscores the value of moving left where practitioners cull data down early, before it reaches a costly review.
- Disposition becomes a discrete phase. Defensibly handling data after a matter closes, whether through deletion, archiving, or a return to normal retention, is now a named obligation rather than an afterthought, a nod to intensifying privacy and security scrutiny.
- Analysis is continuous. Rather than a discrete step, it now runs through every phase as the connective tissue between them—reflecting the growing role of analytics, data science, and AI in decision-making from identification through disposition.
Why this matters beyond the diagram
A reference model is more than an academic picture. It is the shared vocabulary that shapes protocols, budgets, RFPs, and even judicial expectations. Reframing the linear diagram as an iterative, governance-grounded lifecycle validates what leading teams already do and raises the bar for everyone else.
It also aligns tightly with themes we have tracked across our Future of eDiscovery work and our 2025 eDiscovery industry survey: the move upstream toward early case assessment, the embedding of AI throughout the process, and trust as the currency of legal technology. As analysis becomes continuous and AI becomes pervasive, the question is no longer whether to modernize, but how to do so defensibly.
Key considerations for legal and compliance teams
1
Revisit your governance foundation. If IGRM now sits beneath every phase, treat information governance as a discovery control, not a separate program. One should ask: Are retention schedules, data maps, and legal-hold triggers shaping how data is identified and acquired?
2
Shift investment left. The Data Acquisition grouping rewards teams that scope and cull early. Evaluate in-place preservation, indexing, and early case assessment capabilities that reduce volume before it becomes review cost.
3
Operationalize disposition. A defensible end-of-life process for data associated with a matter is now front and center. Document who decides, on what timeline, and how deletion or archiving is evidenced before regulators or opposing counsel ask.
4
Treat analysis as continuous and auditable. If AI and analytics inform decisions at every stage, defensibility depends on documenting how those decisions were made. Build validation, sampling, and audit trails into the workflow, not the post-mortem.
The map is being redrawn. The organizations that treat EDRM 2.0 as more than a diagram, using it to pressure-test governance, acquisition, and disposition against the realities of the AI era, will be the ones that turn a new reference model into a genuine advantage. To discuss what these changes mean for your program, explore KPMG Forensic Technology and eDiscovery capabilities and our KPMG and Relativity alliance.
Meet our team
David Nides
Principal, Advisory, KPMG US
David Nides
Principal, Advisory, KPMG US