What Microsoft Learned About AI at Work
How We Did It
“Every company will have a slightly different approach,” says Nathalie D’Hers, Corporate Vice President of Microsoft Digital, who oversaw the internal rollout to our more than 200,000 employees. “In our case, we zeroed in first on the roles that we knew would gain a lot of benefit.”
It made sense for sales to get first access: After all, they need to know the product inside and out to communicate its value to customers. But beyond that, we found that salespeople are uniquely positioned to benefit from Copilot, whether it’s cutting down on email triage to prioritize leads or gathering relevant info ahead of a client meeting. In early results, our salespeople saved 90 minutes of time per week; 83 percent of them felt they were more productive; and 67 percent said they were able to parlay the time savings into more time with customers.
Next came customer service and support. Nine months ago, they rolled out Copilot to their more than 40,000 support professionals at once, so they could get the entire organization familiar with the technology fast. They had four objectives: reduce time to expertise for agents, streamline access to knowledge, reduce repetitive administrative tasks (to allow people to focus more on customer support, their key priority), and reduce the high volume of inquiries that come in every day.
It’s been a year of learning, but we have started to discover what Copilot can unlock for individual employees and companies as a whole. Most days, it can feel like we’re on a rocket ship. More specifically, like we’re riding on the rocket ship as we’re building it.
The investment has paid off. According to a study last year from our Chief Economist’s office of nearly 10,000 Microsoft support agents, several teams saw, on average, a 12 percent reduction in case handling time and a 10 percent boost in case resolution.
And once HR got access, the department retooled an AI-powered employee resource called Ask HR, which expedited the response time for more complex questions about benefits, payroll, and other HR topics. With HR service advisors using Copilot, employees now get faster and more accurate answers to questions that previously might have taken several days to compile and respond to.
“Our HR service professionals are able to handle employee inquiries more efficiently,” says Kathleen Hogan, Microsoft Executive Vice President and Chief People Officer. “So far we are seeing a 26 percent reduction in initial response time thanks to Copilot.”
From there, we used what we learned from those early adopters to help guide the rollout to the rest of our company.
How You Can Do It Too
Put Copilot where it’s most useful. Whatever department or role you’re targeting, clearly identifying goals before a rollout helps leaders and employees determine from the start what’s working and what’s not. It also helps set appropriate benchmarks for success, whether that’s response times or more effective meetings or other metrics. For guidance, look to our Copilot Scenario Library, which includes suggested use cases and key performance indicators to help orgs determine how Copilot can help.
Go for easy wins too. As you’re going after function-level transformation, use AI to improve simple tasks as well. Gaining confidence and ability early on (for example, asking Copilot to recap a meeting) helps users maintain a healthy growth mindset when they hit the inevitable road bumps.
Give it to entire teams. Rolling out Copilot to entire teams at once—even if they’re small ones—is crucial in promoting peer-to-peer learning: It encourages sharing and learning among the group members, multiplying the impact of the technology. It also allows organizations to see patterns to help identify what’s working (or what’s not).
Make sure to track the impact. To understand how AI is transforming workplace behavior, you’ll need a way to measure its usage. A platform like our Copilot Dashboard can help you plan and measure the impact.