AI Is Creating New ESI. Is Your Discovery Program Ready?
Generative AI has moved from pilot to practice. Legal, compliance, investigations, and business teams now use it to summarize documents, draft content, build timelines, support deposition prep, identify potentially privileged material, and accelerate review. For eDiscovery teams, AI is no longer theoretical; it is part of the electronically stored information (ESI) landscape.
The challenge is readiness. Prompts, responses, uploaded files, summaries, usage logs, and embedded outputs may all become discoverable. Legal teams need to know where this data lives, how long it is retained, whether it can be preserved, and how it can be collected, reviewed, and produced defensibly.
At KPMG LLP, we view AI-assisted data as part of modern ESI, not as a separate discovery category. The practical response is clear: update the data map, understand the tools, document controls, and test workflows before a subpoena, investigation, or regulator creates urgency.
Start with the data map
- Which AI tools are approved, deployed, or embedded?
- What departments have access and are using each tool?
- What AI data is created, stored, logged, or transmitted?
- Where does the AI data reside?
- What retention settings apply for each AI tool?
- Can the AI data be placed on legal hold?
- What export formats preserve context, metadata, and relationships?
Discovery considerations
1
2
3
4
5
6
Keep AI data in context
An AI output often has little meaning without the prompt, source material, and surrounding exchange. Consider treating these as related items or families:
- One prompt and one response
- Source document, prompt, and response
- AI chat thread and referenced documents
- Meeting transcript, AI summary, and user edits
- Generated work product and supporting AI history
KPMG point of view
AI readiness belongs in the broader eDiscovery operating model. KPMG helps organizations assess people, processes, technology, and governance so legal teams can respond faster and more defensibly when AI-related ESI arises.
The takeaway: AI data may be discoverable, and the next wave of AI-driven ESI will expose gaps in programs that have not kept pace. KPMG rapid ediscovery program assessment is designed to help legal teams move from uncertainty to action. In a focused, practical assessment, KPMG can pressure-test the current eDiscovery program across people, process, technology, and governance; evaluate readiness for AI-related data sources; identify gaps in legal hold, preservation, collection, review, and production; and deliver a prioritized roadmap for modernization.
The next step is to use that roadmap to launch targeted workstreams: updating data maps, refining hold and collection protocols, aligning review workflows, and testing AI data scenarios before they arise in an active matter. For teams asking whether they are ready for AI-driven ESI, the rapid ediscovery program assessment is the starting point for building a more defensible, scalable, and future-ready discovery program.
Meet our team