The launch builds on Legal Agent, a contract review and redlining tool Microsoft released to its Frontier early-access program in April. During that preview, more than 1,800 users generated nearly 100,000 edits. Microsoft says users kept roughly seven in ten AI-suggested changes. A new Word feature for navigating defined terms is slated for preview later this month. Microsoft has not yet published independent data on accuracy or on how the product handles client data in sensitive matters, and the Frontier figures come from Microsoft itself.
Microsoft is the fourth major AI company in about five months to ship a dedicated legal offering. Anthropic launched Claude for Legal in May, Google followed with Gemini Enterprise for Legal in August, and OpenAI released Astra for Law in September. All four share a model: legal-specific skills, connectors to specialist legal software, and the promise that lawyers can work without leaving their main platform. Microsoft's distinct angle is distribution. Word is already the default authoring tool for many attorneys, so the product adds capability to an existing workflow rather than asking lawyers to adopt a new app. Microsoft is pitching the product for sensitive legal work, which raises the stakes for accuracy and data protection. Firms evaluating these tools should test tracked-change output and deadline extraction against their own review standards before relying on either, and should confirm how their data is handled under their Microsoft 365 agreements.