Microsoft: Companies Are Going AI-First, Turning to Digital Labor to Scale

More and more, companies are moving decisively toward full-scale AI transformation. The allure of the productivity gains, cost savings, and scalability that AI-powered solutions can deliver is too compelling for business leaders to ignore. In fact, 82 percent say this is the year they’re rethinking their entire strategy and operations to see where AI can fit in. That’s according to Microsoft’s Annual Work Trend Index, which reveals companies no longer dabbling—they’re ready to embed AI into everything they do.

Now in its fifth year, the Work Trend Index offers a global view of how work is evolving. For 2025, the company polled 31,000 workers from across 31 countries. It also included insights from LinkedIn, AI startups, economists, scientists, and academics. Pulling all of the data together, Microsoft says we’re on the verge of a major shift—one where companies embed AI so deeply into their foundations that a whole new kind of business starts to take shape.

“We call it the Frontier Firm,” Colette Stallbaumer, the cofounder of Microsoft’s WorkLab, remarks in a statement. “Every organization’s journey to becoming a frontier firm will be different, but we see it playing out in three phases. In the first phase, AI acts as an assistant, helping people do the same work better and faster. In the next phase, employees start delegating entire projects to agents…And in the final phase, the frontier firm phase, people will manage teams of agents doing work on behalf of an entire team or function.”

Microsoft predicts that within the next two to five years, even companies that haven’t started their AI journey yet will be well on their way to being a so-called Frontier Firm.

In addition to the Work Trend Index, Microsoft is announcing new AI features and agents coming to its Microsoft 365 Copilot app.

The Big AI Shift

For managers to embrace AI, it’s a massive reversal from just a year ago. In 2024, Microsoft’s research showed that, despite employee enthusiasm, companies were slow to adopt the technology. Jared Spataro, Microsoft’s chief marketing officer for AI at Work, explains there’s been a “big secular shift.”

“Last year was about employee-led adoption of AI in the workplace,” he tells me. “This year is about managers starting to really understand how powerful AI can be.” Spataro points out that 81 percent of leaders surveyed said they expect AI agents to be “moderately or extensively integrated” into their company’s AI strategy within the next 12 to 18 months. And some are already taking action: 24 percent of leaders report that they are already deploying AI across their organizations, with 12 percent running pilot programs.

The change in attitude is likely due to two key factors. First, business leaders have seen how rapidly AI technology has advanced in the past year, perhaps surpassing their expectations. They’re realizing that AI is now capable of far more sophisticated and impactful applications. Secondly, the increasing adoption of AI agents in automating various workplace processes and boosting productivity has been a major eye-opener for managers. They’re witnessing firsthand how these bots can streamline operations, reduce costs, and improve efficiency. This is likely sufficient proof of AI’s potential to transform the workplace.

Defining a Frontier Firm

Before we go any further, it’s essential to recognize that when Microsoft says “Frontier Firms,” it’s the equivalent of an organization being “AI-first.”

Spataro admits the terms are interchangeable. “We use the label ‘frontier’ because we liked…[that] frontier models are the way that the industry [refers] to the latest [AI] models. And so we…are thinking about what’s a way that we could articulate AI-first, but that gives people a sense that it’s a bit different this time. It’s not just cloud-first like it was, but there’s something different. And ‘frontier‘ was our label.”

But what makes up a Frontier Firm? He explains that these are organizations that have decided that digital labor is as critical to the business model as human labor. They’ve opted to try to do everything with an AI-first approach.

According to Spataro, five key criteria distinguish those in this group: “The first was [that] the company had to have an organization-wide deployment of AI—we had to be convinced that they had rolled out some AI tools to everyone. Number two, they had to have what we called advanced AI maturity, by looking at indicators like their pace of adoption, innovation, and what they were doing with AI use cases. Number three, they had to be using agents currently. And number four, they had to have additional projected use of agents—we were looking particularly at them using agents. And then number five, they have to have demonstrated that they believe that agents were going to be the key to their ROI for AI.”

I asked him if Microsoft would be considered a Frontier Firm under those criteria. Spataro believes so, remarking that the Windows maker is on the same journey as every other firm and has a mandate from its leadership to be AI-first. “Yeah, I think we are. If we weren’t, we’d be in big trouble. But we are trying to walk the talk, for sure,” he tells me.

Are there industries outside of tech that have embraced the AI-first mindset? Spataro points to consumer packaged goods, financial services, telecommunications, and manufacturing as “high-opportunity” sectors—industries where companies are actively leaning into the technology. In comparison, he lists those in media, automotive, energy, fashion, and entertainment among those lagging, deeming them “higher risks.”

The Journey to the Frontier Firm

That being said, Microsoft has discovered three key shifts that have led to the creation of these Frontier Firms:

Intelligence on Tap

“Intelligence, one of the organization’s most valuable assets, has always been confined within humans,” Alexia Cambon, the senior director at Microsoft Research, explains. “For the first time, we can now access intelligence outside of humans via agents, a new resource that leaders must learn to manage and measure. This is important because human labor is reaching its limits.”

It’s the same philosophy Salesforce has recently been championing—positioning AI-powered digital labor as a supplement to human workers. Spataro acknowledged that Microsoft is not the first to notice this trend, but asserts that what makes its information unique is that the data backs it up. “Sometimes we’re a little bit slower than some, but we’re trying to make sure we can take a comprehensive view,” he proclaimed before adding, “This is what we are quite confident we’ll see because it is what we are actually already seeing out in the wild.”

As Microsoft’s Work Trend Index puts it, intelligence is quickly becoming a durable good—abundant, affordable, and available on demand. In fact, 82 percent of business leaders say they’re confident their companies will turn to digital labor to boost workforce capacity within the next 12 to 18 months. More than half of leaders (53 percent) say more productivity is needed, but 80 percent of the workforce argues that there’s not enough time or energy to do their job. On average, workers find they’re interrupted every two minutes by a meeting, email, or notification.

We “aren’t running out of work. We are running out of human capacity to do it,” Cambon contends. Instead of increasing headcount, organizations are turning to AI to help automate administrative tasks. Shopify, PayPal, United Wholesale Mortgage, and others have started telling employees to use AI first before considering new hires. However, while the Work Trend Index may support such decisions—33 percent of leaders say they’re considering headcount reductions—it’s not a zero-sum game. The data also shows that 78 percent of leaders plan to hire AI experts in the future, with that number rising to 95 percent among Frontier Firms.

AI agents have become so prevalent that they’re the modern-day equivalent of cloud computing, in which they can easily be spun up on demand. Spataro thinks the analogy is an appropriate one, but prefers to call it the elastic workforce, a play on elastic computing. “It really gave you the sense that…it was easy for you to use. You could expand on demand. You could turn it off when you needed to. It really met your needs at the time, and that’s what we have for the first time,” he shares. “Intelligence is something that you’re able to purchase, like electricity or water. You do it independent of hiring an individual—a person—and it gives you more granularity, more flexibility, more control than you’ve ever had.”

Human-Agent Teams Will Upend the Org Chart

The second shift involves establishing hybrid organizations, which are made up of both humans and AI-powered bots. Microsoft predicts the traditional org chart will be replaced by a Work Chart, an outcome-driven model where teams are formed around goals and objectives, not by functions. Companies today are built around domain expertise, such as finance, marketing, and engineering.

Frontier Firms, Cambon believes, will delegate work to two different workforces, human and agentic. “Some functions will become more agent first, and others more human first,” she states, with customer service, marketing, and product development being the top three areas that will receive more AI investment over the next 12 to 18 months. Microsoft’s data shows Frontier Firm workers are more likely than those at non-Frontier Firms to use AI for these tasks: 73 percent vs. 55 percent for marketing, 66 percent vs. 44 percent for customer service, 68 percent vs. 46 percent for internal communications, and 72 percent vs. 54 percent for data science.

When asked why they would assign tasks to AI instead of a human, employers pointed to the technology’s 24/7 availability (42 percent), its speed and quality (30 percent), and ability to generate unlimited ideas on demand (28 percent). As Microsoft notes, these are capabilities humans can’t match. Interestingly, avoiding human traits ranked lowest among the reasons for adopting AI—suggesting that organizations see AI not as a replacement, but as a way to enhance human workers.

And just how many humans and agents there should be on a team will be a “critical business metric” for these transformed businesses. With this human-agent ratio, organizations should be able to measure and optimize the impact of these new teams.

“Companies must think beyond automation, considering when AI and humans together outperform AI alone, when customers value a human touch, and where society expects human accountability,” Cambon says.”

Every Employee Will Be an Agent Boss

In the Frontier Firm era, there will no longer be independent contributors. Everyone will be a manager, just maybe not of other humans. Microsoft believes a new role will emerge, one of an “agent boss.” This is someone who “builds, delegates to, and manages agents to amplify their impact—working smarter, scaling faster, and taking control of their career in the age of AI.”

Under this transformed organization, everyone must start acting like “the CEO of an agent-powered startup” and leverage bots to help with research, data analysis, and other tasks. “Imagine building your own agentic startup within your company, a team of agents designed and customized to your needs,” Cambon explains. “This mandates that all employees learn to manage, especially those just joining the workforce. If it currently takes ten years to become a manager, we will expect the next generation to manage from day one.”

If Microsoft is right, Frontier Firms won’t leave these new managers to figure it out on their own. Nearly half of business leaders (47 percent) say upskilling employees is a top priority over the next 12 to 18 months, and 51 percent see AI training as a core responsibility within the next five years. With AI literacy now the most in-demand skill of the year, according to LinkedIn, investing in worker training has never been more critical.

Last year’s Work Trend Index revealed a grassroots surge in AI adoption across organizations. In 2025, Microsoft claims the dynamic has flipped with business leaders taking the lead. When measured against seven characteristics used to define who is an “agent boss,” managers and executives came out ahead:

• Agent familiarity (67 percent of leaders say they’re familiar or extremely familiar compared to 40 percent of employees)
• Regular AI usage (69 percent of leaders use it often vs. 45 percent of employees)
• Trusting AI for high-stakes work (76 percent of leaders think it’s safe vs. 66 percent of employees)
• Expect to manage agents (36 percent of leaders vs. 21 percent of employees)
• Use AI as a thought leader (54 percent of leaders vs. 41 percent of employees)
• See AI as a career accelerator (79 percent of leaders vs. 67 percent of employees)
• Already seeing AI’s impact (29 percent of leaders vs. 20 percent of employees)

This emerging “agent boss” model is already creating new roles and reshaping hiring plans. The Work Trend Index shows more than a quarter of managers (28 percent) are considering hiring AI workforce managers to run human-agent teams. More than a third (32 percent) are planning to bring on AI agent specialists within the next 12 to 18 months to design, develop, and optimize these tools.

What’s Next?

Spataro describes the 2025 Work Trend Index as a “call to arms.” He concedes that although a new business model has been identified, Microsoft doesn’t have all the answers yet. Nevertheless, the company felt it was important to take stock of what was happening and declare that we’re in a crucial moment. However, Spataro revealed that his team is working on future research, expected to be published this summer, that provides concrete steps for organizations on what to do next.

In the meantime, he suggests managers start by hiring their first digital employee. These aren’t people well-versed in technology, but rather AI agents. Start by establishing the bot’s role and treating it like a human worker. Then, identify a process “that you think is ripe for AI transformation, and really dig into it. Don’t try to boil the ocean, but dig in and get some real results. And don’t do it as a trial. Do it as something that you’re committed to doing.”

Stallbaumer calls 2025 the end of “business as usual” and the beginning of the Frontier Firm era. Microsoft’s Work Trend Index shows that while some skepticism lingers, companies are pressing ahead with practical, grounded AI adoption. The experimentation phase is over—it’s now time to reimagine organizations with AI embedded at their core.

Featured Image: credit: Ken Yeung/Google Imagen 3