iManage

What it is

iManage is the document and email management system that most large law firms and professional-services groups file their work product into. It has been independent since the 2015 management buyout from HP; Bain Capital Tech Opportunities took a minority growth stake in April 2023, and co-founder Neil Araujo is still CEO. As of May 2026 the company reports over 1 million professionals across 4,000 organizations, including 79% of the Am Law 100, 83% of the Global 100, and 40% of the Fortune 100, with 78% of its customer base migrated to iManage Cloud.

The reason iManage belongs on a legal-AI shortlist is not the DMS. It is that iManage has become the permission boundary that legal AI runs through. Harvey, Legora, Litera, Definely, and DraftWise all index it rather than replace it — which makes the question “what can our AI see?” a question about iManage configuration, not about the AI vendor.

Why it shows up in Legal Ops stacks

• The MCP Server turned the repository into an AI-addressable system of record. Announced generally available on 2026-05-14, it replaces a growing list of per-vendor API integrations with one connection that Harvey, Legora, ChatGPT, Claude, Microsoft Copilot, or a firm’s own agents can call. Access stays authenticated, permission-bound, and logged, existing ethical walls hold, and nothing gets bulk-exported to a vendor’s index. It is distributed through Anthropic’s partner ecosystem and the Claude store.
• Security Policy Manager is where the ethical wall actually lives. Need-to-know security at the client and matter level is enforced in iManage, so an AI assistant inherits the wall instead of needing its own parallel model of who may read what.
• Insight+ is the knowledge-search layer over the firm’s own corpus, and it is the surface the MCP Server exposes. AI Enrichment handles document classification and data extraction; Ask iManage is the natural-language assistant for finding and acting on filed work.
• The iManage Work Universal REST v2 API with OAuth 2.0 is what every legal-AI vendor builds its connector against. If a tool claims “DMS integration,” this is usually the API it means.
• At ConnectLive 2026 iManage framed the direction as a context fabric — an architectural layer over content, relationships, and live activity — and added 90 new customer logos during the year.

Pricing

iManage publishes no price. Every product page routes to a demo request, so any number you see comes from resellers and consultants rather than the vendor.

• Reported band: roughly $50–75 per user per month on an approximately 10-seat minimum, per consultant and reseller reporting current as of August 2026. UK channel partners land nearer £30–50 per user per month.
• Implementation is a separate invoice, and a large one: about $10–15K at 10–25 users, $15–25K at 25–50, and $25K+ above 51. A 15-seat firm’s realistic first year is roughly $22–27K all in, against about $11.7K of licensing.
• The DMS number is not the platform number. Insight+, Security Policy Manager, Threat Manager, and Records Manager are licensed separately. A quote that covers Work alone will not cover the governance layer the AI story depends on.

Best for

• Law firms above roughly 50 lawyers, and in-house or accounting teams with a records-retention or ethical-wall obligation, that need matter-centric filing rather than folders.
• The specific case iManage now wins outright: a firm that wants to point Harvey or Legora at twenty years of accumulated work product without exporting that corpus to a vendor’s index, and wants the ethical wall to survive the exercise.

Alternatives

• NetDocuments — the other half of the legal DMS duopoly, cloud-native from the start, reported at $45–65 per user per month for a transactional-grade configuration with the ndMAX AI add-on at a further $15–25. Pick it when you want published per-user AI add-on pricing and have no on-prem legacy to carry.
• Microsoft SharePoint / Microsoft 365 — pick it when you already pay for E3 or E5, your matter volume is low, and nobody is asking you to prove need-to-know access at the matter level. What you give up is the legal metadata model and the ethical-wall enforcement, and you will rebuild both badly.
• DeepJudge — the fastest-growing entrant in the segment, at over 500% year-over-year revenue growth and a $41.2M Series A led by Felicis at roughly $300M, with Freshfields, Holland & Knight, and Cozen O’Connor named. It does not replace the DMS. Pick it alongside iManage when the real problem is search across DMS, email, intranet, and client portals rather than filing.

Watch-outs

• The AI governance controls buyers ask about hardest were roadmap, not shipped. Matter-level restrictions on AI use in Security Policy Manager and AI-agent behavior monitoring in Threat Manager were presented as direction at ConnectLive in May 2026, alongside Insight+ multi-region search, native OCR, and Collaboration Links. Guard: if your rollout plan depends on any of them, get a GA date written into the order form rather than accepting the roadmap slide.
• You cannot price this before a demo, and the second invoice is the surprise. Guard: make the vendor quote licensing, implementation, and each governance module as separate lines, and benchmark the total against a NetDocuments quote for the same seat count — that is the only pricing pressure available to you here.
• MCP grants exactly what the user already has, which cuts both ways. Years of loose need-to-know configuration surface as an AI that returns privileged material to the wrong lawyer; years of over-tight configuration surface as an AI that “can’t find anything.” Guard: audit security classes and NRTA settings on the matters in scope before the first AI pilot, not after the complaint.
• 78% on cloud means 22% are not. On-prem and hybrid deployments do not get the same AI surface on the same schedule. Guard: confirm which deployment your firm is on before you scope any of this, because the answer changes what is actually available to you.
• Being the boundary makes it sticky and expensive. Migration off iManage is a multi-quarter project because every legal-AI tool, add-in, and workflow points at it. Guard: negotiate export format and data-portability terms at the first signature, while you still have room to walk.