Yes, Relativity Is an eDiscovery Tool — But Not for Your Discovery Dump

Yes — Relativity is an eDiscovery tool, full stop. The company describes itself as a "legal data intelligence" platform, and RelativityOne has been rebranded as Relativity aiR, an AI-native platform for litigation, investigations, regulatory inquiries, data breach responses, and FOIA requests. Law firms like Lowenstein Sandler have adopted it as an "AI-powered cloud eDiscovery platform" for complex investigations, litigation, and regulatory matters.

So the headline question has a clear answer. The more useful question for a defender staring at a discovery dump is whether that answer helps you — and it mostly doesn't.

Look at what Relativity is actually built to review. The document types its customers cite are emails, spreadsheets, chats, and PDFs. Its documented use cases are civil litigation, internal corporate investigations, regulatory productions, and FOIA requests. None of the company's own materials mention bodycam video, interrogation footage, jail calls, or 911 audio — the media that makes up the bulk of a modern criminal discovery dump.

That's not a knock on the platform. Relativity serves the DOJ and 198 of the Am Law 200, with 300,000+ users across roughly 40 countries. Its aiR suite is built for document review at scale — finding the five key emails out of five million. Your problem is different: it's not finding a needle in a document haystack, it's making forty hours of bodycam and a two-hour interrogation video reviewable at all.

The mismatch shows up in the workflow:

• eDiscovery platforms ingest text-heavy files; criminal discovery arrives as media — video and audio that need diarized, timestamped transcripts before anything can be searched.
• Civil review workflows end in privilege calls and productions; defense workflows end in cross-exam prep, suppression motions, and Brady/Giglio review.
• eDiscovery citations point to documents; defense work product needs citations to the exact second of footage or page of a police report.

The defense bar has noticed the gap. NACDL's board-approved white paper argues the adversarial system "demands that defenders achieve parity or risk providing representation that falls short of constitutional and ethical standards" — a pointed framing when prosecutors have agency-side platforms and defenders are handed raw media files. And defense attorneys are already reaching for AI: in a survey of 511 US criminal-defense legal professionals, 71% reported using AI in their work and 56% said it outperforms traditional evidence analysis.

Here's the honest caveat: no independent source has tested Relativity against bodycam or jail-call media specifically. The absence of criminal-defense use cases in its own materials is an inference, not a documented limitation. But the structural point holds — a platform built to triage corporate email isn't built to diarize a jail call or sync three camera angles of one arrest.

That's the category Evidence Copilot occupies: a defense-focused evidence workspace where the discovery dump — bodycam, interrogation footage, 911 audio, scanned police reports — gets transcribed, OCR'd, and cited down to the exact moment, with every AI answer marked for human review. Different evidence, different workflow, different tool.

What Relativity Actually Does: Document Review at Scale

If you've ever opened a 40-hour discovery dump and wondered whether the "eDiscovery platform" your civil-litigation colleagues rave about would help — the short answer is that Relativity solves a different problem. Here's what it actually does.

RelativityOne, rebranded in September 2026 as Relativity aiR, is a legal data intelligence platform for litigation, internal investigations, regulatory inquiries, data breach responses, and FOIA requests. Firms like Lowenstein Sandler use it to review emails, spreadsheets, chats, and PDFs at massive scale — not bodycam video, jail calls, or police reports.

Its documented functions center on document review:

• Large-volume review — vendor-reported results include a government agency reviewing 650,000 documents in one week, with key documents found in roughly 20% of the usual time
• AI review and privilege — the aiR suite became generally available in late 2024, with customers reporting up to 70% cost savings and 80% less review time (vendor-reported)
• Government work — aiR for Review launched in RelativityOne Government as the only FedRAMP-authorized generative AI built for document review
• Regulatory and data-subject productions — vendor-reported precision as high as 99.8%, with some productions finished in 48 hours

Note the caveats: those numbers come from Relativity's own announcements, not independent verification. A 99.8% precision rate on a regulatory production tells you nothing about how a platform handles a three-hour interrogation video.

Here's the honest part: on human oversight, Relativity and defense-side tools agree. Relativity's AI Principles state that "humans have clarity and control over every component of the AI" — the same human-in-the-loop posture that NACDL's Parity in Practice white paper demands of defenders. Human review is table stakes across legal AI now, not a differentiator.

What actually differs is the evidence and the workflow. Relativity's documented use cases never mention bodycam footage, 911 audio, jail calls, or report-vs-video impeachment — the daily work of a criminal-defense practice. A defense workspace like Evidence Copilot is built around exactly that: timestamped transcripts with clickable citations into the video, OCR'd scanned reports, and auto-flagged inconsistencies between the police report and the bodycam.

So yes — Relativity is an eDiscovery tool, and a formidable one. It's just built for a discovery world of documents, not one where the key evidence talks back at 30 frames per second.

The Defense-Side Gap: Why Defenders Need a Different Tool Category

Here's the uncomfortable math: the prosecution's side of the aisle runs on purpose-built platforms, while most defenders are still scrubbing through bodycam files named X13qrt9__x3455 on a media player that won't even speed up. NACDL puts the stakes plainly in its white paper: the adversarial system "demands that defenders achieve parity or risk providing representation that falls short of constitutional and ethical standards." Parity isn't about having a tool — it's about having the right category of tool.

Relativity is not that category. It is, by its own description and its customers', an AI-powered cloud eDiscovery platform built for litigation, internal investigations, regulatory inquiries, data breach responses, and FOIA requests. The document types it's documented to handle are emails, spreadsheets, chats, and PDFs. None of Relativity's published materials mention bodycam video, interrogation footage, jail calls, 911 audio, or police reports — the actual contents of a criminal discovery dump.

The defense bar already knows AI belongs in the workflow. In a survey of 511 US criminal-defense legal professionals, 56% said AI outperforms traditional methods for evidence analysis, and 71% have already used AI in their work. But 94% harbor at least one concern, led by output accuracy at 63%. That combination — high demand, high skepticism — is exactly why a litigation eDiscovery platform doesn't answer the question. Defenders don't need "AI that reviews documents." They need AI whose every answer can be checked against the source.

That's what defines the defense-evidence category as distinct from eDiscovery:

• Media transcription with diarization — bodycam, interrogation room footage, and jail calls turned into speaker-labeled, word-timestamped transcripts, not treated as unreviewable attachments.
• Redaction for discovery turnover — frame-accurate video and word-level audio redaction that produces derived files while originals stay untouched, with an audit trail.
• Timestamped citations — every AI answer links to the exact second of video or page of the PDF, so a skeptical attorney can verify before relying on it.
• Report-vs-bodycam inconsistency review — automated signals when the police report's narrative diverges from what the footage actually shows, surfacing impeachment material before anyone thinks to ask.

This is the workspace Evidence Copilot was built around: a grounded evidence-review environment where the discovery dump goes in and timestamped, citable, human-reviewable work product comes out — with every AI output marked as pending attorney review.

Notably, even Relativity embraces human control, stating its AI is designed so "humans have clarity and control over every component." Human review is now table stakes across legal AI. What differs is the evidence itself — and that's where defenders need a tool built for their side of the courtroom, not the civil-litigation one.

Evaluating Evidence Tools for Criminal Defense: A Practical Checklist

Before you sign up for any platform, run it through the same screen you'd apply to an expert witness: does it actually do the defense work you need, and can you verify everything it produces? That screen matters more now that 71% of criminal-defense legal professionals have used AI in their work — and 94% report at least one concern about it, led by output accuracy.

If you've been evaluating eDiscovery platforms like Relativity, note what its own materials emphasize: document review at scale for litigation, investigations, regulatory inquiries, and FOIA — emails, spreadsheets, chats, PDFs. Bodycam, jail calls, and interrogation footage aren't in that workflow. So build your checklist around the media-heavy reality of criminal discovery:

• Media transcription with word-level timestamps. Ask specifically about bodycam, jail calls, and 911 audio — with speaker diarization and one-click speaker renaming. A transcript you can't pin to the exact second is useless for impeachment.
• OCR for scanned police reports. Most discovery PDFs have no text layer. The tool should make them searchable and citable, page by page.
• Redaction that never touches originals. Frame-accurate video redaction, word-level audio muting, and PDF redaction — every output a new derived file, every action audited. This is non-negotiable for discovery turnover.
• Citations to the exact moment. Every AI answer should link to the precise video second or PDF page it came from. If it can't, you can't verify it — and 63% of defense attorneys cite output accuracy as their top AI concern.
• Human review built in. Every AI output should be marked pending review, per the oversight guidance in NACDL's white paper — and the tool should never make legal conclusions for you.

One more check: inconsistency detection. The best tools surface report-vs-bodycam and statement-vs-call-audio conflicts on ingest, before you think to ask — the task 26% of defense attorneys already use AI for.

Evidence Copilot was built around this exact checklist: grounded citations into the exact second of video or page of PDF, auto-signals for candidate inconsistencies, redaction with immutable originals, and every AI output marked for human review. Pricing is published — Solo at $39/user/mo, PD Office at $750/attorney/yr — with no feature tiers and a hard cap on overage so runaway intake can't overbill you.

If you want to see how it handles your own discovery dump, book a 20-minute walkthrough at evidencecopilot.com/book, or start on the Solo plan at evidencecopilot.com/signup and test it on a real case.