Atlas Fuse

DMS vs Knowledge Management for Law Firms | AtlasFuse

Written by Product Director | Knowledge Management, Search, AI, Governance | Sep 30, 2026, 8:48:42 AM

At a few legal technology events this year - including ILTACON 2026 and, most recently, Alternative Legal IT 2026 - we have had versions of the same conversation with IT leaders.

Many are understandably wary of taking content outside the boundaries of the document management system. The concern is usually expressed in a simple phrase:

"We want everything to stay in iManage."

Or NetDocuments. The product changes; the principle is the same.

It is an understandable position. The document management system is trusted, secure and already embedded in the way lawyers work. It contains the client files, matter documents, emails, precedents and work product accumulated over many years. Keeping knowledge there can feel like the safest way to preserve permissions, ethical walls and a single source of truth.

The discussions at those events also made something else clear: this is rarely resistance to knowledge management itself. It is a legitimate concern about duplication, governance and creating another repository for lawyers to navigate.

My answer is that this is not really a debate about where documents should be stored. It is about the difference between matter content and firm knowledge. Those are related, but they are not the same thing.

The short answer:

Law firms should continue to use their DMS as the system of record for client and matter work. Modern platforms such as iManage Insight+ can also provide meaningful curation, metadata, knowledge collections and lifecycle workflows within the DMS. The wider strategic question is whether the firm's complete knowledge strategy should be centered there, particularly when so much operational knowledge and day-to-day work sits across Microsoft 365 and other systems.

The better model is not "DMS or knowledge platform". It is a connected model in which the DMS remains authoritative for matter-centric records, while the firm deliberately decides where reusable knowledge should be curated, governed and activated.

A document is not automatically knowledge

A DMS is very good at answering questions such as:

  • Which matter does this document belong to?
  • Who created or edited it?
  • Which version is the latest?
  • Who is permitted to access it?
  • How should it be retained as part of the client file?

Those are important questions. They are not, however, the only questions a lawyer - or an AI assistant - needs to answer before reusing something.

They also need to know:

  • Is this the firm's preferred precedent or simply a document that was used once?
  • Has it been reviewed and approved for wider use?
  • Is the legal or commercial position still current?
  • In which jurisdictions and circumstances should it be used?
  • Does it reflect the firm's normal position or an unusual client concession?
  • Who owns it, and when will it next be reviewed?
  • What related guidance, expertise or policy should be considered with it?

This leads to an important distinction: Permission is not endorsement. Search relevance is not authority.

A lawyer may be entitled to see a document without the firm intending them to rely on it as a precedent. An AI system can retrieve a highly relevant document that is nevertheless outdated, exceptional or unsuitable for the task.

Solutions such as iManage Insight+ are specifically designed to help address this through enriched metadata, curated knowledge collections and publishing workflows. That is an important part of the answer. It does not remove the need for the firm to decide what constitutes authoritative knowledge, who owns it and where it should be made available.

That distinction matters more - not less - as firms introduce Microsoft Copilot, legal AI platforms and intelligent agents.

Matter content and firm knowledge serve different purposes

Matter content

Firm knowledge

Created while delivering client work

Selected because it has value beyond one matter

Organized around clients and matters

Organized around practices, topics, jurisdictions and business purposes

May include drafts, duplicates and negotiated exceptions

Curated to represent an approved or useful firm position

Access is governed by clients and matter permissions

Reuse is governed by authority, audience, purpose and policy

Retained according to records obligations

Maintained according to continuing value and currency

Explain what happened on a matter

Helps the firm act consistently on future matters

 

The distinction is not absolute. Valuable knowledge frequently begins as matter content. The question is what happens next.

If a particularly strong clause, argument, checklist or piece of advice remains buried in a completed matter, it is technically stored - but it has not necessarily become institutional knowledge.

Why "keep everything in the DMS" is so appealing

Firms usually have good reasons for taking this position. They trust the DMS security model. Lawyers already use it. Ethical walls and matter permissions are established. The firm wants to avoid duplication, additional repositories and yet another place people must remember to search.

Those concerns should not be dismissed. Nor should anyone suggest moving every client and matter document into a separate knowledge platform. The DMS remains essential.

But keeping every document in one system does not necessarily produce one coherent body of knowledge. In practice, it can produce a very large body of content in which the firm's best knowledge is difficult to distinguish from everything surrounding it.

That creates a new risk for AI. If an assistant is given access to thousands or millions of permissioned documents, how does it know which represent the firm's current, preferred and approved position? More content is not always better grounding.

Modern DMS platforms now have a credible KM story

This is not an argument that DMS providers are standing still. Far from it.

iManage Insight+ should be recognized for what it is: a knowledge search and management solution, not simply an extension of basic DMS search. It supports enriched metadata, curated knowledge collections and workflows through which knowledge teams can submit, enrich, publish, maintain and archive knowledge assets. It also inherits iManage permissions and ethical walls and can make approved knowledge available to Ask iManage and external AI tools through MCP.

Please also see: iManage Insight+ alternative for law firms, legal knowledge management 

NetDocuments is following a similarly ambitious direction with its Legal Context Graph, designed to map relationships between matters, documents, communications and people and make that context available to AI agents.

These are important developments. They also mean the discussion can no longer be reduced to "the DMS stores documents; the knowledge platform does search". So, the balanced view is not that a DMS cannot manage knowledge. It plainly can - and for firms committed to keeping curated legal work products inside the DMS, Insight+ offers a coherent approach. The more useful questions are:

Where does the firm define and govern the knowledge it actively endorses for reuse - and does that knowledge begin and end with content held in the DMS?

For many firms, the knowledge landscape is wider. It includes policies and guidance in SharePoint, collaboration in Teams, expertise and people information, intranet content, learning material, client and sector knowledge, business data and approved work product from the DMS.

The architectural decision is therefore about more than where a precedent is stored. It is about where the firm wants to create a common knowledge experience across those sources and where it wants governance, lifecycle and AI activation to be coordinated.

Where the Microsoft 365-native position stands out

This is where AtlasFuse presents a different center of gravity. AtlasFuse is native to Microsoft 365 and operates within the customer's Microsoft environment. For firms that already use SharePoint, Teams, Outlook, Microsoft Copilot and the wider Microsoft platform, this allows knowledge management to become part of the digital workplace rather than a destination attached primarily to the DMS.

That matters in several practical ways:

  • Knowledge can span systems. AtlasFuse can connect approved work products from iManage with policies, guidance, expertise, intranet content and information held elsewhere.
  • The existing Microsoft governance investment can be used. Knowledge held within Microsoft 365 can participate in the firm's established identity, permissions and governance model, including Microsoft Entra ID and, where configured, Microsoft Purview sensitivity, retention and data protection controls.
  • Knowledge is available in the flow of work. The same governed knowledge can be surfaced through SharePoint, Teams, Outlook, enterprise search and Microsoft Copilot rather than requiring a separate knowledge destination.
  • The firm is not tied to one AI experience. Governed knowledge can support Copilot, specialist legal AI tools, assistants and agents, subject to the appropriate controls.
  • The wider Microsoft platform remains available. Firms can use existing capabilities such as Power Automate and Copilot Studio to extend workflows and AI experiences without rebuilding every process in a separate technology stack.

This does not mean content must simply be copied out of iManage. AtlasFuse can work with references, connected search and selectively promoted knowledge assets, allowing the DMS to remain the source of record while Microsoft 365 provides the broader knowledge and activation layer.


A better model: promote knowledge rather than migrate documents

The most practical approach is a knowledge promotion model.

Illustrative AtlasFuse knowledge promotion model

1. Work remains in the DMS

Drafts, correspondence, transaction documents, filings and matter records continue to be created and governed in iManage or NetDocuments.

2. Valuable know-how is identified

The firm identifies high-value clauses, precedents, arguments, checklists, templates, lessons learned and other material with potential value beyond the original matter.

3. Selected material is promoted where it can be governed and reused

The item is turned into a genuine knowledge asset. That may mean curating it within a DMS solution such as iManage or creating a matter neutral asset in AtlasFuse with practical commentary, jurisdiction, provenance and appropriate approval. The right choice depends on the asset, its audience and where it needs to be consumed.

4. The knowledge asset is governed

The promoted knowledge receives an owner, audience, status, review date, provenance and lifecycle. Superseded material can be withdrawn without relying on every lawyer - or every AI tool - to recognize that it is no longer appropriate.

5. The knowledge is activated

The same trusted asset can be used through enterprise search, Microsoft Copilot, AI assistants, legal agents, intranets and business workflows.

This does not have to mean copying every source document. Depending on the information architecture and risk requirements, the promoted asset could be:

  • an approved template or precedent managed in AtlasFuse;

  • a matter-neutral derivative of an original document;
  • a practical knowledge page explaining when and how to use something;
  • a governed reference back to the source in the DMS; or
  • a dynamic collection combining selected DMS content with policies, expertise and information from other systems.
The important change is that the firm - not simply the location of the file - defines what the asset means and how it should be used.

 

Why this matters for legal AI

Legal AI platforms are increasingly good at retrieving and reasoning over large volumes of information. But reliable retrieval still depends on the quality of the knowledge supplied to them.

Without a governed knowledge layer, an AI assistant may find:

  • several conflicting versions of a clause;
  • a bespoke position negotiated for one client;
  • advice based on an outdated regulatory position;
  • a document that is factually relevant but not approved for reuse;
  • material without the explanatory context a lawyer would apply instinctively.

This is not necessarily a model failure. It is often a knowledge-management failure.

The firm needs a way to tell its AI systems: these are the sources we trust; this is the context in which they apply; this is who owns them; and this is how current they are.

Crucially, that knowledge should not become captive to a single AI application. A firm may use Microsoft Copilot, a specialist legal AI platform and its own agents at the same time. Each should be able to draw on the same governed knowledge foundation, subject to the same permissions and policies.

The role of AtlasFuse

AtlasFuse works alongside iManage and other enterprise systems from its native position within Microsoft 365. The DMS can remain the authoritative system of record for the matter - and Insight+ may manage curated knowledge that the firm deliberately wants to retain there - while AtlasFuse creates a broader, connected knowledge layer across the organization.

That provides a layer through which the firm can:

  • distinguish authoritative knowledge from the wider document corpus;
  • add consistent metadata and business context across systems;
  • manage ownership, provenance, approval and review;
  • create trusted knowledge collections for particular practices, topics or purposes;
  • connect knowledge with relevant people and expertise; and
  • make the same governed knowledge available to search, Copilot, assistants, agents and workflows.

Because AtlasFuse is embedded in Microsoft 365, firms can also build on capabilities they already license and govern: Microsoft Entra ID for identity, SharePoint and Teams for collaboration, Microsoft Purview for applicable information-protection and lifecycle policies, and Power Automate or Copilot Studio for workflow and agentic use cases.

This is not about replacing the DMS or dismissing Insight+. It is about ensuring that the firm's knowledge strategy is not constrained by the boundaries of any one repository - and that the knowledge generated through client work can be combined safely with the wider organizational context needed by people and AI.

Questions every law firm should ask

If your firm currently expects the DMS to manage all institutional knowledge, consider the following:

  1. Can lawyers clearly distinguish approved precedents from merely similar documents?
  2. Does every important knowledge asset have an owner and review date?
  3. Can the firm withdraw superseded knowledge from every AI and search experience consistently?
  4. Can knowledge be organized independently of the client or matter that produced it?
  5. Can DMS content be combined with policies, expertise and information held elsewhere?
  6. Do AI tools know which sources the firm endorses, or only which sources a user can access?
  7. Can the firm reuse the same governed knowledge across multiple AI platforms?
  8. Does the proposed approach govern knowledge outside the DMS as effectively as knowledge inside it?
  9. Can the firm build on its existing Microsoft 365, Entra ID, Purview, Power Automate and Copilot investments?

If the answers are unclear, the issue is not that the DMS has failed. It may be that DMS centered architecture does not cover the firm's complete knowledge estate.

From stored work to institutional knowledge

Law firms have spent years building secure, well-governed repositories of client and matter content. That investment remains fundamental. The next challenge is to turn the best of that work into knowledge the wider firm can find, trust and reuse - and that AI can consume safely.

The goal should not be to move everything out of the DMS. Nor should it be to treat everything inside the DMS as knowledge. It should be to create a deliberate path from work product to institutional knowledge:

The DMS records and increasingly curates the work the firm has done. AtlasFuse connects that work with the wider organization, governs knowledge within Microsoft 365 and puts it to work across people, search and AI.

That is the distinction that will determine whether legal AI simply searches for more documents or genuinely operates from the firm's collective knowledge.

FAQ

What is the difference between a legal DMS and a knowledge management platform?

A legal document management system manages documents and emails associated with clients and matters, including versioning, access, security and records obligations. A knowledge management platform helps the firm identify, enrich, approve, maintain and reuse authoritative knowledge across matters, systems and working environments.

Should law firms move documents out of iManage or NetDocuments?

Not as a general rule. Client and matter records should normally remain governed by the DMS. Selected material with wider value can then be promoted into reusable firm knowledge, either as a governed reference, a curated copy, a matter neutral derivative or a knowledge asset with additional context.

Can iManage or NetDocuments support knowledge management?

Yes. iManage Insight+ explicitly provides knowledge curation, enriched metadata, curated collections, publishing and archiving workflows, governed discovery and AI access. NetDocuments is also extending its platform through its Legal Context Graph. The question for each firm is not whether these products can support KM, but whether a DMS centered model covers knowledge across Microsoft 365 and other systems and delivers it consistently to every required user and AI experience.

Why does curated knowledge improve AI answers?

Curated knowledge gives AI clearer authority signals and richer context. It helps prevent drafts, outdated material and exceptional matter-specific positions from being treated as standard firm guidance simply because they are relevant and accessible.

How does AtlasFuse work with a legal DMS?

AtlasFuse works alongside systems such as iManage rather than replacing them. From within Microsoft 365, it connects selected DMS content with SharePoint, Teams, expertise and other enterprise sources, then adds the structure, governance and context needed to make firm knowledge discoverable, reusable and suitable for search, Copilot, assistants and agents.