
Last updated: 21 November 2026. This guide uses the EDRM model as published by EDRM, the Federal Rules of Civil Procedure and Federal Rules of Evidence as published by Cornell’s Legal Information Institute, Practice Direction 57AD for the Business and Property Courts of England and Wales, and vendors’ own product pages and announcements.
Short answer: eDiscovery software helps legal teams find, preserve, collect, process, review and produce electronically stored information (ESI) for litigation and investigations. Most teams map tools to the nine stages of the Electronic Discovery Reference Model (EDRM). No single product covers every stage equally well, so firms usually combine a review platform with hold, collection and forensic tools.
Key takeaways
- The EDRM model has nine stages, from information governance to presentation. EDRM itself describes the model as iterative, not a linear waterfall.
- Review platforms such as Relativity, Everlaw, DISCO, Reveal, Nuix Discover, Casepoint and Nextpoint cover processing to production. Holds, collection and forensics often need separate tools.
- In US federal court, Rules 26(b)(1), 26(f) and 37(e) set the duties your software has to support: proportionate discovery, an agreed ESI plan and reasonable preservation steps.
- Pricing models differ more than features. Per-gigabyte hosting, per-user plans and AI add-ons produce very different totals on the same matter, and few vendors publish prices.
What is eDiscovery software?
eDiscovery software is any tool used to handle ESI for disclosure or discovery: emails, chat messages, shared drive files, phone data and the metadata that goes with them. The term covers everything from a legal hold notice system to a cloud review platform that runs technology assisted review (TAR) across millions of documents.
The market is crowded and consolidating. In August 2023 Reveal bought both Logikcull and IPRO, and the ComplexDiscovery deal tracker counted 16 eDiscovery deals in 2026 up to late October, among them Relativity’s purchase of Gavel in June and Repario’s acquisition of UnitedLex in September. Product names and owners change often, so check the vendor’s current page before relying on any list, including this one.
If you are newer to the subject, our guide to running a legal hold without spoliation covers the duty that starts the whole process, and our piece on eDiscovery and AI explains what generative AI did and did not change in review.
What are the EDRM stages, and which tools fit each?
The Electronic Discovery Reference Model was founded in 2005 by George Socha and Tom Gelbmann. Duke Law School acquired it in 2016, and in October 2019 Mary Mack and Kaylee Walstad bought it from Duke’s Bolch Judicial Institute. The EDRM model calls itself a “conceptual view of the e-discovery process, not a literal, linear or waterfall model”. In practice you loop back: a new custodian sends you back to identification, preservation and collection.
| EDRM stage | What happens | Typical tool types | Examples |
|---|---|---|---|
| Information governance | Managing data from creation to disposal so that discovery costs less when it comes | Retention and records tools in the email and document systems | Microsoft Purview, Google Vault |
| Identification | Finding where relevant ESI sits and how much is in scope | Data mapping, custodian interviews, early case assessment | Purview and Vault search, Hanzo (collaboration data) |
| Preservation | Stopping deletion or alteration of relevant ESI | Legal hold notices and in-place holds | Exterro Legal Hold and Preservation, ZDiscovery, Reveal Hold, Purview, Vault |
| Collection | Gathering ESI in a forensically sound way | Forensic imaging, mobile extraction, cloud connectors | Cellebrite, Magnet Forensics, Exterro FTK Forensic Toolkit |
| Processing | Reducing volume and converting data into a reviewable form | De-duplication, de-NISTing, text extraction, threading | Nuix, Relativity, Reveal |
| Review | Coding documents for relevance and privilege | Review platforms, TAR, generative AI review | Relativity, Everlaw, DISCO, Reveal, Logikcull, Casepoint, Nextpoint |
| Analysis | Finding patterns, people, topics and timelines | Analytics, clustering, chronology builders, AI chat | The same review platforms |
| Production | Delivering ESI to the other side in an agreed form | Redaction, Bates numbering, load files, privilege logs | The same review platforms |
| Presentation | Showing ESI at depositions, hearings and trial | Trial presentation and deposition tools | Trial Director (Reveal), Nextpoint |
The examples are illustrations based on each vendor’s own product pages in November 2026, not recommendations. Our eDiscovery and litigation software directory lists the main review platforms with short descriptions.
Which tools handle governance, preservation and collection?
Most corporate data now sits in Microsoft 365 or Google Workspace, and both have built-in discovery tools that many teams use before anything else.
- Microsoft Purview eDiscovery. Places holds on Exchange mailboxes, SharePoint sites, OneDrive accounts and Teams, searches and collects in place, copies results into review sets and exports in native or EDRM format for third-party review tools. Microsoft says it retired all classic eDiscovery experiences on 31 August 2025, so older guides that describe classic Content Search are out of date. See the Purview eDiscovery overview.
- Google Vault. Sets retention rules, places holds on accounts, organisational units and groups, and searches and exports Gmail, Drive, Calendar, Chat, Meet recordings, Groups, Sites and Gemini app messages. Google says that from 1 November 2025 admins need a Vault licence to keep using it.
- Dedicated legal hold tools. Exterro, Zapproved’s ZDiscovery, Reveal Hold and Casepoint send custodian notices, track acknowledgements and keep the audit trail a court may ask to see.
- Collaboration data. Slack and Teams messages do not behave like email. Hanzo sells preservation and collection for collaboration platforms, and the large review vendors have added their own connectors.
- Forensic collection. For phones, laptops and anything where metadata may be contested, firms use forensic tools such as Cellebrite, Magnet Forensics or Exterro’s FTK Forensic Toolkit, usually through a specialist examiner rather than the case team.
Built-in tools are cheap and fast, but they only see their own system. Once a matter involves phones, third-party apps or a former employee’s personal account, you will need something else.

Which tools handle processing, review and analysis?
This is where most of the spend goes and where products differ most. The main platforms are:
- Relativity. RelativityOne is its cloud platform. aiR for Review, its generative AI review tool, became generally available in 2024, and in October 2025 Relativity said aiR for Review and aiR for Privilege would be included in RelativityOne’s standard pricing and packaging.
- Everlaw. A cloud platform for review, analytics and case building. At ILTACON in September 2026 it announced integrations with Thomson Reuters CoCounsel Legal, Harvey, Google’s Gemini Enterprise for Legal and Microsoft 365 Copilot, positioning itself as the system of record for evidence.
- DISCO. Cloud review with an AI review feature, Auto Review, which it updated in September 2026 so users can refine AI tagging through plain-language feedback.
- Reveal. Owns Logikcull (aimed at smaller teams), Reveal Enterprise, Onna, Reveal Hold and Trial Director, and launched an agentic AI suite, Reveal AI, in 2026.
- Nuix Discover. Known for high-volume processing. Its AI Chat feature is due for general availability in December 2026.
- Casepoint and Nextpoint. Casepoint covers eDiscovery, legal hold and FOIA work and has a strong government customer base. Nextpoint combines review with deposition and trial preparation tools aimed at smaller litigation teams.
Machine-assisted review is not new. In Da Silva Moore v. Publicis Groupe (S.D.N.Y., 24 February 2012), Magistrate Judge Andrew Peck wrote that “computer-assisted review now can be considered judicially-approved for use in appropriate cases”. In England, Master Matthews approved predictive coding for a population of over 3 million documents in Pyrrho Investments v MWB Property [2016] EWHC 256 (Ch). Generative AI review is newer and less tested, so validation matters more, not less. Our guide to AI document review covers how these tools work and what they cost, and the AI hallucination sanctions record shows what happens when lawyers file AI output without checking it.
What about production and presentation?
Production is where disputes about form arise. Under Federal Rule of Civil Procedure 34(b)(2)(E), if a request does not specify a form for ESI, a party must produce it in a form in which it is ordinarily maintained or in a reasonably usable form, and need not produce the same ESI in more than one form. Your platform needs to produce native files, images and load files in whatever form the ESI protocol requires, apply redactions that cannot be reversed and generate a privilege log.
Ask for a Federal Rule of Evidence 502(d) order early. It lets the court order that privilege is not waived by disclosure connected with the case, which protects you if the review tool or the reviewers miss a privileged document. Presentation tools such as Trial Director and Nextpoint’s presentation mode then put exhibits and deposition clips in front of the judge or jury.
What do the court rules expect your software to support?
United States. FRCP 26(b)(1) limits discovery to nonprivileged matter relevant to a claim or defence and “proportional to the needs of the case”. Rule 26(b)(2)(B) lets a party resist ESI that is “not reasonably accessible because of undue burden or cost”, and Rule 26(f)(3)(C) requires the discovery plan to address preservation and the form of production of ESI. Our guide to FRCP Rule 16 explains how these issues end up in the scheduling order.
FRCP 37(e), rewritten in 2015, applies when ESI that should have been preserved is lost because a party failed to take “reasonable steps to preserve it” and it cannot be restored or replaced. On a finding of prejudice the court may order measures “no greater than necessary to cure the prejudice”. Adverse inference instructions, dismissal or default judgment need a finding of intent to deprive. Hold tools with an audit trail are how you prove reasonable steps.
England and Wales. In the Business and Property Courts, Practice Direction 57AD requires legal representatives to cooperate to promote reliable, efficient and cost-effective disclosure, “including through the use of technology”. Paragraph 9.6 expects parties to discuss “software or analytical tools, including technology assisted review software and techniques”, and paragraph 4 requires them to suspend deletion and notify relevant employees in writing. A Disclosure Review Working Group is reviewing PD 57AD. Its July 2026 summary lists clarifying how AI fits within TAR as a theme and says it hopes to consult on recommendations late in 2026 or early in 2027.
How is eDiscovery software priced?
Few vendors publish prices, so expect to ask for a quote. The models you will meet are:
- Per gigabyte per month. The traditional model, charged on data hosted. Logikcull’s Pay As You Go plan, for example, charges per GB each month with a 10 GB minimum, though its pricing page gives no dollar figure.
- Per user with unlimited data. Nextpoint advertises per-user pricing with “No surprise fees. No processing charges. No hosting costs.”
- Per matter or annual subscription. Common for corporate legal teams with steady caseloads.
- AI as an add-on or as standard. Some vendors charge separately for generative AI review. Relativity moved its aiR review tools into standard RelativityOne packaging.
Model the total on a real matter: data in, data after processing, months hosted after review ends, user numbers and the cost of exporting a production set if you later leave. Billing for review work also needs care; our explainer on UTBMS codes explains the L300 discovery task codes, such as L320 for document production, that many clients require.
How do you choose eDiscovery software?
- Start with your data sources. List where your clients’ data actually sits: Microsoft 365, Google Workspace, Slack, phones, legacy file shares. Rule out tools that cannot reach them.
- Decide who will run it. Enterprise platforms assume a litigation support team. If your case team will run it, test it with them.
- Check validation reporting. For TAR or AI review, you need recall and precision figures, sampling records and a written description of the method in case it is challenged.
- Ask about hosting and data residency. Cross-border matters may need data held in a particular country or region.
- Test production. Run a small production in the form your usual ESI protocol requires before you sign.
- Plan the exit. Find out what it costs to export your data and work product when the matter ends or you switch vendor.
Many firms below a certain size never buy a platform and use a service provider instead, which moves the selection question to the provider’s technology and validation practice.
This article is general information about US federal practice and the law of England and Wales as of 21 November 2026, not legal advice or a product recommendation. Check the rules of your court and each vendor’s current terms before relying on them.
Frequently asked questions
What is the EDRM model?
The Electronic Discovery Reference Model is a framework, founded in 2005, that divides eDiscovery into nine stages: information governance, identification, preservation, collection, processing, review, analysis, production and presentation. EDRM describes it as iterative rather than a linear process.
What is the most widely used eDiscovery software?
No independent, published market share figures were verified for this guide. Relativity, Everlaw, DISCO, Reveal, Nuix, Casepoint and Nextpoint are the platforms most often named, and Microsoft Purview and Google Vault are widely used for holds and early collection.
Is Microsoft Purview enough for eDiscovery?
For data inside Microsoft 365 it handles holds, search, review sets and export, and it can export in EDRM format to other review tools. It cannot reach phones, Slack or other outside systems, so larger matters usually need further tools.
Do courts accept AI or predictive coding in document review?
Yes, for TAR. A US federal court approved computer-assisted review in Da Silva Moore in 2012, and the English High Court did so in Pyrrho in 2016. Generative AI review is newer, so courts expect a documented and validated process.
What happens if ESI is lost during litigation?
In US federal court, Rule 37(e) lets the judge order measures to cure prejudice if a party failed to take reasonable preservation steps. Adverse inference instructions, dismissal or default require a finding of intent to deprive the other side of the information.
How much does eDiscovery software cost?
It depends on the pricing model and data volume. Vendors charge per gigabyte hosted, per user, per matter or by subscription, and few publish prices. Ask for a written quote based on a matter of the size you actually run.
Where the legal industry reads first.
Enjoyed this article? Get the biggest legal industry updates, deals, appointments, insights and expert interviews in your inbox, free.
No spam. Unsubscribe anytime.From Legal Desire
Is your firm the one being cited, or the one being skipped?
We ran the test. On several everyday legal questions, software companies are answering and law firms are not. We help firms fix that, and we build the technology behind the practice.
Lawyers and firms: get featured on Legal Desire
Share your deals, lateral moves, firm news and insights with 1M+ monthly readers: founders, law firm partners, general counsel and practising lawyers across the US, UK, EU and GCC.
Get featuredSubmit a press release

