As you probably already know, contract review AI has existed in specialist platforms for a few years now.

What Microsoft has built is a version of that capability that lives inside Word rather than in a separate application.

The core functions: it analyzes contracts clause by clause against your internal playbooks, flags risks and non-standard provisions, generates negotiation-ready redlines using tracked changes, and compares document versions to surface what changed and what the implications are.

An early testing group reported drafting contracts roughly 60% faster.

Citations in the tool link directly to the underlying text, so a reviewing attorney can verify each suggestion in seconds without leaving the document.

The objection I hear is usually some version of: "We have Copilot. Nobody uses it."

Fair.

The old Copilot in Word was a general-purpose drafting and summarizing tool - useful for some things, but not purpose-built for legal work.

This is different.

Microsoft built the Legal Agent with the Robin AI team - a contract AI specialist they acquired specifically for this product.

It uses what they describe as a "deterministic resolution layer" rather than pure generative AI, meaning it applies structured rules to document edits rather than simply predicting text.

That distinction matters when a single word change in a contract carries real legal risk.

THE REAL PICTURE

Look, as I see it, the limitations here are as important as the capabilities.

Microsoft's own documentation is clear: the Legal Agent does not provide legal advice, does not make professional determinations, and is not a substitute for the judgment of a qualified attorney.

Human review of every suggested change is required before implementation.

The Frontier early-access program is US-only and rolling out gradually - most firms will not have broad access yet.

And the honest word from a legal operations director quoted in Legal IT Insider on May 1st: "The functionality isn't blowing anyone away who's already across the specialist tools."

That is the balanced picture.

First, a clarification: Harvey and Legora are not just contract review tools.

They handle research, workflow automation, matter analysis, and knowledge management - full legal AI platforms priced for enterprise use.

Neither publishes pricing, but customer reports and analyst estimates converge on roughly $1,000 to $1,200 per attorney per month with 20-seat minimums - around $240,000 to $290,000 per year for a starter deployment.

For firms already on those platforms, the Microsoft Legal Agent is additive.

For mid-size and smaller firms that cannot justify that scale, Microsoft just raised the floor for contract review specifically, inside the tool your team already uses, at no marginal cost if you have M365 Copilot.

That is the real story here.

This is where most rollouts will go wrong.

The Legal Agent is only as useful as the contract playbooks your firm has built.

If your firm has not documented its standard positions on key clauses - liability caps, indemnity language, IP ownership, termination rights - the tool has nothing to check against.

Most firms have not done this work.

The second issue is privilege.

Configuration must happen before a single client document goes through the system.

Default M365 tenant settings were not designed with attorney-client privilege in mind, and a misconfiguration does not announce itself.

"Installed" is the easy part.

"Working, safely, with client documents" is a project.

Here is what I would suggest this week.

First, ask IT whether your firm is enrolled in Microsoft's Frontier program for M365 Copilot - if you are, some users may already have access.

Second, start documenting your contract playbook: your firm's standard positions on the clauses that appear in every deal.

You need this before the feature becomes useful.

Third, have privilege configuration reviewed before any client document goes through the system.

If your team needs help building those playbooks, that is exactly the kind of focused session we run at Cyberaktive.

We built a free interactive tool that weighs your firm size, use cases, and budget against the major legal AI platforms AND custom build options - and gives you a personalized recommendation with the reasoning behind it.

How is your firm primarily handling AI-powered document review right now?

A) We're using Microsoft Copilot or the new Legal Agent in Word

B) We use a dedicated legal AI tool (Harvey, Legora, Spellbook, or similar)

C) General AI tools like ChatGPT or Claude, without legal-specific configuration

D) Still manual - we have not deployed AI for document review yet

[Reply with your letter choice] - I'll share the results in the next edition.

The best AI tool is the one your attorneys actually use.

Microsoft just made legal AI as frictionless as opening a Word document - and that is genuinely significant.

But "frictionless access" and "correctly configured for privilege" are not the same thing, and the gap between them is where I expect most firms to fumble the next six months.

The firms that get real value from the Legal Agent will be the ones that treated the rollout as a project - playbooks built, privilege reviewed, team trained - before any client file went through it.

Hit reply and let me know: is your firm testing the Legal Agent?

What is working, and what is not?

Liam Barnes

Want help building your contract playbook and configuring Microsoft Copilot the right way for legal?

Grab some time to chat

(if you don’t see a suitable time, just shoot me an email [email protected])

How Did We Do?

Your feedback shapes what comes next.
Let us know if this edition hit the mark or missed.

Too vague? Too detailed? Too long? Too Short? Too pink?

👍 Loved it | 👎 Not for me | 💡 I Have a Suggestion

OR: [email protected]

Was this week’s newsletter forwarded to you?

Sign up, it’s free.