How AI Will Transform Work: The Next Measure of Enterprise AI Success
Microsoft 365 Copilot just crossed 30 million paid users, with new signups more than doubling quarter-over-quarter. That's impressive on its own. But here's what really matters: across every industry, AI is shifting from helpful assistant to active participant. It's becoming woven into how work actually happens—not just another tool people occasionally use. Organizations are rearchitecting their entire workflows around it. Employees are directing autonomous agents to own entire tasks, projects, and processes, expanding what their teams can accomplish and amplifying their impact.
The first wave of enterprise AI was about adoption metrics: deployments, licenses, hours saved. Those benefits will keep multiplying. The next wave will be defined by transformation—empowering every person and every organization on the planet to achieve more. The opportunity is growing as humans and AI agents collaborate to build things that previously seemed impossible.
New data from Microsoft 365 Copilot's usage signals, along with real examples from customers worldwide and Microsoft's own teams, shows this shift is already happening. It shows up in three ways: how deeply people are integrating AI into their work, what AI can actually accomplish—from working faster to unlocking entirely new capabilities—and how organizations are customizing it for specific roles. Let's dig into each, starting with the technology itself.
A Breakthrough in What's Possible
Over the past few months, a new category of AI product has emerged: a model wrapped in an agent system that can close its own feedback loop—planning, executing, checking, and fixing work to deliver final results. This approach enabled Copilot Cowork, which rolled out broadly in June. Cowork takes on multi-step jobs: you define the task, and it runs from start to finish and hands back a complete result.
Its multi-model design lets you run whichever models a task needs, so capabilities expand alongside new models. Testing shows Cowork costs 30-40% less than single-model alternatives. Microsoft also released its first autonomous agent, Microsoft Scout, which operates in the background with its own identity and permissions, keeping work moving even when you're focused elsewhere. All outputs are built on Work IQ to understand the context shaping work in your organization—roles, projects, priorities, people's institutional knowledge—so they fit uniquely to your business.
What's striking is the leap forward not just in products shipped to customers, but in how engineering teams build them. Despite many people contributing to Cowork, the core team stayed intentionally small—growing from 3 engineers initially to just 9 at broad release. Yet the speed to market is remarkable. Cowork went from concept to preview to adoption by half of Fortune 500 companies in just 6 months. The breakthrough came from rebuilding the software development process around agents, letting a small team achieve velocity, scale, and automation that traditional engineering workflows can't match.
Microsoft Scout started similarly—a small team working alongside autonomous agents, sharing context, feedback, and clear specs—and again, they moved from initial concept to a working product in weeks instead of months. "Previously, I spent a lot of time helping teams scale: resizing large groups, coordinating work, making sure everything moved forward together," says Omar Shahine, Corporate Vice President of Microsoft Scout. "Now what matters is enabling expert teams to move fast, experiment safely, and continuously capture what they learn so each project goes further than the last." Teams like these—lean, focused, shaped by human-agent collaboration—signal how knowledge work will transform in the coming months and years.
The numbers tell the story: Copilot conversations have nearly doubled year-over-year. Weekly engagement with Copilot now matches Outlook and Teams. And 49% of Cowork workflows involve multiple steps.
How Deep and How Wide AI Usage Goes
As AI products mature and the methods for building them improve, adoption within organizations is getting deeper too. In just the past year, conversations per user have nearly doubled, and the number of users interacting with multiple Copilot features has tripled. The fact that average weekly engagement with Copilot now equals Outlook and Teams? That's a signal AI is becoming part of the daily work rhythm.
When you look at what people ask Copilot Cowork to do most in any given week, you see a rise in multi-step requests—things like analyzing a dataset and drafting an email explaining it, or researching a problem and writing code to solve it. When you account for these multi-step workflows, analysis-related work jumps to 49% of total tasks, up from 29% when analysis was treated as standalone. What's interesting here is that data shows AI is moving beyond helping people finish individual tasks to helping them complete entire workflows.
Looking across industries, traditional leaders like manufacturing, banking, and software/tech remain among the highest adopters of agents in the Microsoft 365 ecosystem. But what's surprising is that automotive and healthcare are emerging as the fastest-growing agent adopters, suggesting the value proposition is spreading well beyond early-mover industries.
The clearest sign of momentum? How fast organizations are scaling Copilot across their workforce. The number of customers with over 50,000 licenses has grown more than 7x year-over-year. Meanwhile, enterprise customers deploying Copilot to most of their staff have increased nearly 75% quarter-over-quarter—a signal of how mission-critical it's become to operations. Most striking: the time to reach high adoption (active monthly usage above 80% across the entire user base) has dropped from months to just days over the past year. These shifts show organizations are rapidly embedding AI and agents into daily workflows.
What AI Can Actually Do
At the most basic level, AI-powered speed creates efficiency. But organizations seeing real value are pushing past efficiency-for-efficiency's-sake and discovering how AI fundamentally changes the work itself—and can expand what an organization is capable of.
Microsoft's Cloud Supply Chain team took a "simplify before you automate" approach. They mapped and streamlined 6 end-to-end workflows, built a shared data foundation, then deployed over 70 custom-built agents across planning, sourcing, order fulfillment, and logistics. Result: 75% reduction in cycle time on selected workflows. The lesson: fix the work before automating it, not the other way around.
Cycle time is one measure of progress. Decision quality is another. At Eaton, a smart energy management company, root cause analysis had relied on time-consuming manual investigations. To lighten the load, they deployed an AI-backed quality agent that helps experts review production quality reports to quickly surface multiple potential root causes, trends, and supplier insights. Applied to roughly 5,000 reports so far, the solution has accelerated analysis, strengthened data-driven decision-making, enabled earlier risk identification and preventive action—all while keeping Eaton's expert team at the center of every call.
Other organizations are discovering what's possible by empowering people closest to the problem to build solutions. Premera Blue Cross set out to let employees create their own AI agents to reduce repetitive manual work and clear growing backlogs and manual processes. The health insurer started by weaving Microsoft 365 Copilot and Copilot Studio into tools employees already use daily, backed by training and clear guidelines so staff could experiment without risking member trust. Since then, the team has built over 900 agents and counting—including a contract assistant that uses intelligent document processing to shrink manual workflows from 30-45 minutes down to about 3 minutes. That frees staff to focus on members needing care or answers to pressing questions.
Going further, AI can unlock capabilities that weren't previously possible. At Levi Strauss & Co., where AI is woven across the entire business from consumer insights to HR, an automated system analyzed 1,100 standard financial operating procedures and catalogued 18,000 individual tasks in a single day—revealing work-level detail that would've been impossible to surface otherwise. "The automation identified an opportunity I didn't think could be quantified," says Lisa Sterling, VP of Finance at Levi's.
And at EY, a global consulting firm, an Autonomous Sourcing Agent (ASA) automates low-cost, low-risk transactions across the entire procurement workflow—from requisition and negotiation through purchase order creation. It can negotiate with one or multiple vendors simultaneously on cost, delivery time, and warranty terms, while keeping humans in place to confirm and resolve issues. Since pilot testing in October 2025, the ASA has supported over 200 transactions and is expected to handle 1,500 next year. The result: faster, more efficient procurement that helps teams extract more value from suppliers while keeping important decisions in human hands.
Building Expertise at Scale
The difference between organizations seeing transformation and those still waiting isn't how widely they've deployed AI—it's how well they've tailored it to their actual work. Most organizations still default to a generic model: give everyone the same tool and let them figure it out. The organizations creating differentiated value are doing something different. They start from how specific roles actually work, then build AI around those patterns.
S&P Global shows how AI can give more people access to deep expertise by putting trusted information directly into workflows. As demand for AI-ready data grew, the company used Microsoft Copilot Agent Builder and Microsoft Graph connectors to create a Copilot agent in Microsoft Teams, giving customers a simpler way to access and interact with their commodity research and market insights. By delivering data extraction 95% faster and comparative analysis 98% faster, the solution helps customers turn massive information volumes into actionable insight more quickly, enabling faster, better-informed decisions across global energy and commodity markets.
Microsoft's sales team segmented different seller personas into specialized AI agents—a research agent, an analytics agent, a sales agent, a deals agent, and MSXi Copilot (a sales reporting and intelligence gatherer)—each designed for specific moments in how sellers spend their time. With agents handling more administrative and coordination work, seller time spent on direct customer engagement doubled from 25% to 50%. The pilot team saw average revenue per seller jump 9.4% and deal velocity speed up 20%, and reported a 3x increase in adoption of priority use cases in a follow-up survey.
Kantar took a grassroots approach. At this data and marketing analytics company, a "people agent" for HR questions—really a system of about 10 agents with 60 topics and related flows working together—is being built and scaled in collaboration with the people who actually handle and support these requests across the entire People team. The system is resolving an increasing share without human intervention: an employment confirmation letter that used to take three days now gets answered in 120 seconds. Kantar reports that the percentage of HR requests handled this way has grown from zero in February to roughly 40%, with a year-end target of 95%.
Microsoft's People Operations team handles over a million employee interactions annually across 100+ systems, doing work that truly matters to people. They're using Frontier Tuning to scale HR expertise across entire processes and employee moments rather than improving individual tasks, capturing expert knowledge as tasks, skills, and criteria that define quality standards. The result is an AI-powered system that can own operational work with the context, standards, and organizational knowledge needed to succeed—converting expert secrets into organizational capability. The team can see real-world performance and tune factors—tools, context, skills, even the model itself—to ensure consistent results. Tools built with Frontier Tuning, now deployed in People Operations, deliver 5x better task completion success rates compared to basic multi-purpose models. As routine operations are increasingly handled by AI, Microsoft's HR experts can focus on complex employee interactions, nuanced decisions, and strategic work where human understanding matters most.
What Comes Next
As AI technology matures and organizations worldwide begin seeing real value emerge, the metrics that matter are shifting too. We'll see greater focus on depth of use, measurable changes in the work itself, tool development and governance, redesigned roles and workflows, and the emergence of entirely new capabilities. Organizations building sustainable advantage are those where technology, operating models, and human capability are evolving together. That's the portrait of the Frontier Firm taking shape.
Description: Microsoft 365 Copilot hits 30M paid users. But the real story isn't adoption rates—it's how AI is reshaping jobs and workflows across industries.
No Comment to " How AI Will Transform Work: The Next Measure of Enterprise AI Success "