Epiq

What it is

Epiq is one of the large legal-services and eDiscovery providers. It sells its own review platform, managed document review, court reporting, legal staffing, and breach response to corporations, law firms, and courts. The company says it has about 4,000 people in 17 countries. OMERS Private Equity and Harvest Partners took it private in 2016 and merged it with DTI, and David Dobson was CEO as of November 2025.

Epiq is reshaping itself as a technology company, mostly by buying:

• November 2025 — sold its business-process outsourcing division (Epiq GBTS) to K2 Services, saying it would focus on “services-enabled technology” and SaaS.
• 24 March 2026 — bought LitLingo, AI monitoring of employee communications for antitrust, bribery, and employment risk.
• 9 July 2026 — bought Tenor Legal, folded into its Epiq Counsel flexible-lawyer business.
• 25 September 2026 — bought Canopy, the breach-response data-mining software it had partnered with since at least 2025, now part of Epiq AI.

This page covers the part most legal ops teams actually buy: Epiq Discover (renamed from Epiq Discovery) and the Epiq AI layer on top of it. On 19 February 2026 Epiq folded its AI Discovery Assistant into Discover. That gives you Epiq Assist (formerly Epiq Chat) for fact questions, timelines, and witness fact sheets with links back to source documents, plus automated AI review that tags documents for issues, relevance, privilege, and PII. Epiq Narrate (formerly EpiqTMX) handles testimony and exhibits after review. Data can go on to review in Discover or in Relativity.

The question it answers

Epiq, Lighthouse, and Consilio are the three names on most outsourced-review panels, and buyers usually compare them on bid price alone. Epiq’s difference is that it publishes a per-document price for its AI, which most rivals do not. That turns “should AI do first-level review on this matter?” into arithmetic you can do before you sign.

The performance figures are all Epiq’s own: Assist does fact research “up to 45 times faster” than manual methods, the platform analyzes up to 500,000 documents an hour, more than 80% of review is automated, and review finishes up to 90% faster than TAR or linear review. Baker Donelson is the named law-firm user. None of these numbers has been independently tested. Use them to decide what to pilot, not to set a budget.

Pricing reality

Epiq publishes more pricing than most of this segment, in two public listings:

• US, AWS Marketplace: Epiq Discover Core SaaS subscription, 1 TB, $45,000 for 12 months, with $6/GB/month overage. That works out to about $3.75/GB/month if you fill the terabyte. Other volumes go through a private offer.
• UK, G-Cloud 14 (valid from 5 March 2026): £270 per matter per month including 50 GB, then £6.00/GB/month hosting, £1.10/GB/month dormant hosting (£40 minimum), project manager £195/hour, litigation support £155/hour.
• Epiq AI (same UK schedule): £0.37 per document for AI for Review plus Assist, or £0.12 per document for Assist only. AI consulting runs £155–£350/hour.

The AI fee is a one-time charge on every document ingested into an Epiq AI collection, not on the documents a person later reviews. At those rates a 500,000-document collection costs £185,000 for AI review plus Assist, or £60,000 for Assist only. Culling before ingestion is where the savings are. Managed human review and US AI rates are quote-only.

Best for

In-house litigation, investigations, and privacy leads at large companies, and AmLaw litigation-support teams, who want one vendor for platform, AI review, reviewers, and testimony across many matters a year, and who want to model AI cost per document before committing. Epiq is also the strongest of the three large providers when a data breach turns into a notification project, since it now owns the Canopy software as well as the review team.

Not for you if you run a handful of small matters a year, or your team wants full control of its own review platform. A self-serve tool will be cheaper and simpler.

Alternatives — and when to pick them instead

• Relativity — the market-share leader in review platforms. Pick RelativityOne with aiR directly when your litigation-support team runs and defends its own review and wants to buy services matter by matter.
• Lighthouse — the closest managed-review rival. Pick it when privilege review is the bottleneck and you want its published IQ Priv results behind the bid.
• Consilio — the third large provider, with its own Aurora AI suite. Get it as the comparison quote on any Epiq or Lighthouse bid.
• Everlaw — the cloud-native challenger with generative AI built into the platform. Pick it for self-serve review with predictable pricing across many mid-size matters.
• DISCO or Reveal — pick these when you want self-serve AI review from a software-first vendor, with services optional.
• Logikcull — pick it for small matters and subpoena response.

Watch-outs

• The AI fee applies to everything you ingest. Loading a lightly culled collection into Epiq AI multiplies the bill at £0.37 a document. Guard: run date, custodian, and deduplication culling in Early Data Assessment first, and put a document cap per matter in the SOW that needs written approval to exceed.
• Every speed figure is a vendor number. The 45x, 80%, and 90% claims come from Epiq’s own release. Guard: require sampled recall and precision on your own matter before production, and tie the AI-review fee to meeting a set threshold.
• An acquisition run means integration risk. Four deals or divestitures in eleven months, plus renamed products, means features and contracts are still moving. Guard: name the products, data locations, and support terms in the master agreement, and add a right to export data and coding history if a product is renamed, merged, or retired.
• No public API or MCP server. Discover and Epiq AI are reached through Epiq’s own interface and Service Cloud, so agents and matter-management tools cannot pull status directly. Guard: make productions, coding exports, and privilege logs in editable formats a contracted deliverable at matter close.