Is the 70/30 Rule for AI True? What It Means for Law Firms

Is the 70/30 Rule for AI True? What It Means for Law Firms

Is the 70/30 Rule for AI True? What It Means for Law Firms

Legal AI and Document Context

Is the 70/30 Rule for AI True? What It Means for Law Firms

AI is changing how organizations handle repetitive work. It can summarise documents, compare contracts, extract information, and organise large volumes of content. However, the quality of its output depends on the information it can access.

A document may contain the right information but still be unsuitable for a task because it is outdated, belongs to a different matter or has access restrictions. Without this context, AI can produce answers that require additional checking and correction.

The opportunity is not simply to automate more work. It is to give AI access to the right documents, the latest information, and the context needed to use them correctly.

Is the 70/30 rule a proven measure of AI automation?

The idea that AI can handle 70% of routine work while humans focus on the remaining 30% is a useful way to describe the potential of automation. However, it is not a proven industry standard, and the actual split varies by task and organisation.

McKinsey Global Institute estimated in November 2025 that demonstrated technologies could theoretically automate activities accounting for around 57% of US work hours. This represents technical potential, not a prediction of how much work AI will perform.

The amount of work AI can take depends partly on the information available to it. Even a capable system may struggle when documents are outdated, disconnected, or missing crucial details.

Why document context matters

In legal work, a document cannot always be understood alone. Its relevance depends on the client, matter, version, status, and access permissions associated with it.

Much of this information is maintained in a document management system (DMS). When documents are downloaded and shared with AI separately, their associated context may not be able to travel with them.

This creates several problems:

  • Outdated information: AI may use an old draft instead of the latest document.

  • Repeated work: Lawyers must locate files, download them, and provide them to AI again.

  • Additional verification: Lawyers spend time checking document versions and confirming whether the information is still valid.

  • Missing context: AI may not have the matter details or document relationships needed to identify relevant information.

The result is more manual work around AI, reducing the time it was intended to save.

How connecting the DMS to AI changes the process

Connecting AI directly to the DMS provides a way to address this problem at its source.

Instead of repeatedly downloading documents and uploading them into separate AI tools, users can work with documents from their managed content environment. Document context and updates can remain available through the connection, depending on the integration and its synchronization settings.

Figure 2

Documents that arrive with their context

7030
Connected at the source, each document reaches AI with its context attached. What is left for people is judgment, not re-sharing and re-checking.

This changes the workflow in three important ways.

  1. 1. Access to current documents

    When document updates are reflected through the connection, AI can work with newer information instead of relying on copies that may have become outdated. This reduces the effort required to locate and re-share revised files.

  2. 2. Context that stays with the document

    Matter details, document status, and other available metadata help establish what a document relates to. Preserving this information gives AI a stronger basis for identifying relevant material.

  3. 3. Less repetitive preparation and checking

    Reducing manual downloads, uploads, and repeated file selection can simplify everyday workflows. Lawyers can spend less time preparing information for AI and checking its source, leaving more time for substantive review.

A connected DMS does not guarantee that every answer will be correct. The quality of the underlying data, access controls, retrieval process, and human review still matter.

Can better context reduce the remaining 30%?

Better context can reduce some of the manual work that remains around AI, but there is no verified percentage showing how much it can eliminate.

Tasks such as finding the right document, confirming its status, and checking whether a newer version exists can often be reduced through better integration and document management.

Legal interpretation, risk assessment, negotiation, and advice still require professional judgment. The goal is to reduce avoidable work around these activities, not to remove the responsibility for making decisions.

This distinction matters. AI becomes more useful when it can work from relevant, current information without requiring people to repeatedly prepare and verify the same material.

Building the right foundation for AI

A reliable AI workflow needs more access to documents. It needs a dependable connection to the information those documents represent.

Organizations should ensure that their approach:

  • Connects AI to the DMS rather than relying on disconnected file copies.

  • Keeps document updates and relevant metadata available through the integration.

  • Preserves appropriate access permissions and confidentiality controls.

  • Helps AI identify the correct document and version.

  • Supports source checking and human review where required.

The aim is to make current, relevant information available when AI needs it, while keeping the DMS as the system of record.

Connecting the DMS to AI can reduce repeated downloads, file sharing, document preparation, and checks for outdated information. It can also help preserve the context needed to identify relevant documents.

Frequently Asked Questions

The idea that AI can handle 70% of routine work while humans focus on the remaining 30% is a useful way to describe the potential of automation. However, it is not a proven industry standard, and the actual split varies by task and organisation.

In legal work, a document cannot always be understood alone. Its relevance depends on the client, matter, version, status, and access permissions associated with it.

Connecting the DMS to AI can reduce repeated downloads, file sharing, document preparation, and checks for outdated information. It can also help preserve the context needed to identify relevant documents.

Better context can reduce some of the manual work that remains around AI, but there is no verified percentage showing how much it can eliminate. Legal interpretation, risk assessment, negotiation, and advice still require professional judgment.

Source: McKinsey Global Institute, Agents, robots, and us: Skill partnerships in the age of AI (November 2025).