Virginia Briggs, the chief executive of one of Australia’s largest law firms, doesn’t start her day without consulting Microsoft 365 Copilot. By her own account, she deploys the AI across “every task, every day”—from briefing herself ahead of client meetings to polishing internal town hall messages. And she’s not alone. According to a new Microsoft case study, half of MinterEllison’s early Copilot users are now saving two to five hours daily, and one-fifth are reclaiming five hours or more. For the 2,500-person firm, that’s a potential injection of thousands of hours per month into higher-value work, and a glimpse into how AI could reshape the business of law.

The AI That’s On Every Lawyer’s Desktop

Briggs’s tool of choice is Microsoft 365 Copilot, but it’s the more advanced “Researcher” agent that she credits with the biggest productivity leaps. Copilot itself embeds generative AI inside Word, PowerPoint, Outlook, and Teams, and can summarize email threads, generate meeting transcripts, or spin up first drafts of documents. But Researcher goes deeper: it’s designed to handle multi-step research tasks, scouring both internal firm data and the open web, then producing a cited, editable report. For a CEO who might need to understand a new client’s business overnight or churn through a mountain of pre-meeting materials, that capability turns a daunting research sprint into a prompt and a few minutes.

The firm’s deployment isn’t one-size-fits-all. Alongside off-the-shelf Copilot, MinterEllison built internal generative AI tools hosted in its own Microsoft Azure environment. These “Chat with ME” services handle particularly sensitive drafting—such as initial versions of legal letters—where data must never leave the firm’s controlled cloud. The blend of standard Copilot, advanced agents, and bespoke secure tools is what makes the MinterEllison story more than just an anecdote; it’s a template for how enterprises can bolt AI onto existing processes without ripping out their security playbook.

The Verdict: 2 to 5 Hours Saved Daily

The numbers coming out of the pilot are eye-catching. In a survey of early users, MinterEllison and Microsoft reported that half were saving between two and five hours each day, and 20% were saving more than five hours. User satisfaction was sky-high: 89% found the tools intuitive, and 90% said they’d recommend them to a colleague. Those figures aren’t audited or longitudinal, but they line up with other enterprise rollouts where document-heavy workflows see dramatic early gains.

How are lawyers spending those reclaimed hours? The case study points to several high-value shifts: senior partners are devoting more time to client strategy rather than drafting, juniors are getting faster feedback loops, and the entire firm is accelerating its turnaround on urgent matters. For a sector where billable hours are king, that efficiency creates a conundrum—but one that MinterEllison seems willing to tackle head-on.

The Researcher Agent: A Closer Look

Researcher is not just a souped-up chatbot. It’s an agent that asks clarifying questions, then pulls together information from work emails, shared files, calendar entries, and the broader web to build a structured, cited report. Microsoft has designed it to integrate with third-party connectors like Salesforce or ServiceNow where authorized, meaning a lawyer could task Researcher with compiling a briefing on a client’s contractual history, current market standing, and potential litigation risks—all in one pass. Briggs has described using it to generate meeting scripts and potential collaboration angles, making her appear exceptionally well-prepared to external executives. That kind of just-in-time intelligence is transforming executive decision-making.

From Pilot to Mandate: The 80% Rule

MinterEllison isn’t tiptoeing into AI. It publicly set a target that 80% of its staff should be using AI at least weekly by early 2025—a mandate that puts it ahead of most legal competitors. The rollout combines executive sponsorship, role-based training, internal “Champions” on Microsoft Teams who share tips, and constant feedback loops. This top-down pressure, backed by enablement, is straight out of the digital transformation playbook. It also sends a clear signal: AI literacy is no longer optional for lawyers at this firm.

If you’re a lawyer or paralegal, expect your daily routine to shift. Routine drafting, meeting notes, and email triage will increasingly be handled by AI, freeing you for client counseling, strategy, and complex problem-solving. But you’ll need to become a skilled editor and validator of AI outputs. For IT administrators, the lesson is clear: set tenant controls early. Configure transcription policies in Teams, decide which agents users can access, and consider a secure internal environment for the most sensitive tasks. And for partners evaluating pricing models, the rise of AI may force a move away from billable hours toward value-based fees—a long-predicted shift that AI could accelerate.

Protecting Client Secrets When AI Is Everywhere

No law firm can afford to let client data leak, and the use of generative AI raises legitimate fears. MinterEllison’s approach leans heavily on Microsoft’s tenant controls: Copilot only sees data the logged-in user is already allowed to access, and admins can toggle exactly when AI is permitted in Teams meetings—or even require transcripts for auditing. On top of that, the firm’s internal Azure-hosted tools create an extra sandbox where sensitive work can proceed without touching the public internet.

Still, the human element is the real firewall. Every AI output must be reviewed by a qualified lawyer before it leaves the building. That human-in-the-loop rule is now table stakes, but it only works if the user is diligent. Past incidents elsewhere—where lawyers submitted AI-hallucinated case citations to courts—serve as a stark reminder of what happens when the loop breaks. MinterEllison’s culture of validation is as important as any technical guardrail.

How Your Firm Can Follow the Same Path

MinterEllison’s story offers a replicable playbook. First, pilot Copilot in discrete, high-volume areas: start with meeting transcription, email summarization, and first-draft document creation. These tasks deliver quick wins with minimal risk. Second, lock down governance from day one. Configure tenant policies to restrict which agents can access which data, and if you’re dealing with ultra-sensitive material, consider hosting your own secure instance in Azure or a private cloud. Third, treat AI as a junior associate: its work must always be double-checked and refined by a human with skin in the game. Sign-off protocols and digital provenance trails aren’t just nice-to-haves; they’re essential for professional liability.

Fourth, don’t ignore training. MinterEllison’s 80% weekly usage target is ambitious precisely because it’s backed by role-based coaching and internal champions who troubleshoot in real time. If your firm is serious about adoption, you’ll need to invest in change management, not just technology. Finally, measure more than time saved—track client outcomes, quality of work, and career development for juniors to ensure AI doesn’t stunt professional growth.

The Bigger Picture: What Comes After the Pilot

MinterEllison is hardly the only firm experimenting with AI, but its CEO’s full-throated, daily-use endorsement gives it unusual momentum. Microsoft is already rolling out even more capable agents—like Analyst, which performs deep data analysis—and integrating Copilot with third-party connectors. The next 12 months will likely see a proliferation of legal-specific AI tools, and MinterEllison’s internal environment could serve as a testing ground. The firm’s experience also puts pressure on law schools and bar associations to update ethical guidelines: if a machine can draft a near-perfect contract, what does “competence” mean for a newly minted lawyer? These are questions the profession can no longer afford to ignore.

For now, Virginia Briggs and her colleagues have shown that AI, when embraced at the top and shored up by rigorous governance, can move the needle not just on productivity but on the very culture of a law firm. The old model of grinding out documents in isolation is giving way to a faster, more collaborative dance with an AI partner. Whether that dance ultimately leads to happier lawyers and better client outcomes depends on how thoughtfully the humans lead.