Microsoft is a global company, with more than 200,000 employees working in offices around the world. The working conditions and rights of those employees are governed by the laws that apply to that country or region.
In parts of Europe and elsewhere, relationships with our employees are governed by works councils. These councils play a major role in vetting and approving any new technology that might impact our workers and their jobs.
At Microsoft Digital, the company’s IT organization, we have gained a ton of insight by working closely with our works councils, especially as we all embrace the rise of generative AI and new tools like Microsoft 365 Copilot.
“With the speed of AI innovation today, we can’t sit and wait. We need our internal users, including our works council members, to be forward-thinking early adopters and help us drive AI transformation.”
Irina Chemerys, regional experience lead, Microsoft Digital
This experience means we are able to help guide customers who are dealing with their own works councils—many of which also have questions and concerns about the agentic workplace of the future, which is coming fast.
As we continue on our journey to becoming an AI-first Frontier Firm, we are collaborating closely with our works councils to make sure that we address their concerns and follow all applicable regulations. This process also helps us make better products that meet the evolving needs of our customers, wherever they are.
“With the speed of AI innovation today, we can’t sit and wait,” says Chemerys, a regional experience lead who oversees works council engagements within Microsoft Digital. “We need our internal users, including our works council members, to be forward-thinking early adopters and help us drive AI transformation.”
How works councils work
Our works councils serve as the voice of our employees in some geographies (especially in Europe), advocating for their rights and interests within the workplace. Typically, they have purview over topics like workplace health and safety, pay and benefits, hiring, business reorganizations, training, and more.
As AI technology becomes increasingly commonplace across many industries, our works councils—along with all of our employees—are at the forefront of the complex discussions regarding the implications of AI for the modern workforce.
While our relationship with Microsoft works councils has always been cooperative and collaborative, how we engage with them for product reviews has evolved over time. What used to be somewhat impromptu or inconsistent engagements have changed to become more strategic and programmatic opportunities for feedback, which can end up greatly improving our products.
Chemerys helped lead the Microsoft Digital effort to streamline the approval process for new technology across works council countries. She drove the development of a global solution that uses a single request form and platform for our works councils worldwide, helping them communicate with Microsoft Digital, product groups, legal, HR, and others at the company.
This simplified communication across the board, and facilitated collaboration among all works councils, allowing smaller countries to take advantage of resources from larger ones and creating a more cohesive community. The unified approach significantly improved coordination, collaboration, and, importantly, trust among works councils.
“Trust is more essential than anything else in terms of collaborating with works councils effectively, especially in the context of AI,” Chemerys says. “To build that trust, you need transparency. And the way you build transparency is by having a well-documented and effective request process.”
Approving Copilot: The tolerance phase
Trust and good communication were linchpins of the process we used to gain works council approval for Microsoft 365 Copilot. Considering how new generative AI tools are, and the widely promoted fears surrounding their potential impact on the workforce, there were understandable concerns raised by some of our works councils about Copilot.
Germany was one country where our works councils were particularly wary of AI. They raised questions about the ways that Copilot could be used to evaluate individual employee performance or make impermissible inferences about individual employees without the data to support them. For example, Copilot might be asked to generate a ranking of employee performance during a meeting, something that fell outside the boundaries of our Microsoft responsible AI principles.
“The earliest versions of these generative AI tools lacked guardrails,” says Carsten Schleicher, chairman of the Microsoft works council in Germany. “You could ask them anything and get an answer back—even questions about religion, race, gender, ethnicity, etc. There were also concerns about AI tools generating false information—so-called ‘hallucinations.’”
“AI is in the world; if you deny your employees access to it, you’ll fall behind. It was absolutely necessary to find a constructive way to deal with AI, and to use it in a fair and transparent way in our company.”
Carsten Schleicher, chairman, Microsoft German Works Council
Faced with these concerns, but also wanting to get as much feedback on Copilot from our European employees as possible, we decided to introduce a tolerance phase.
Countries such as Germany, France, and the Netherlands were included in this approach, which allowed for controlled deployment of the tool while still enabling employees to try it out.
“As a works council, it’s your goal to protect the employees, but you also want the company to be successful,” Schleicher says. “AI is in the world; if you deny your employees access to it, you’ll fall behind. It was absolutely necessary to find a constructive way to deal with AI, and to use it in a fair and transparent way in our company.”
Some members of our works councils were part of the first Copilot deployment wave, and their feedback was then channeled back to the product engineering team. This early access helped the councils quickly reach an agreement that deployment of Copilot could continue, while also leading to product improvements that benefited all our customers.
The tolerance phased ended in the spring of 2025, and Copilot is now approved for use by Microsoft employees worldwide.
Getting ready for the agentic future
After Copilot was approved, the next challenge for our works councils was the world of AI agents. As a Frontier Firm, Microsoft is gearing up for a workplace where employees are routinely aided in their work by digital agents. Eventually, agents may become our “digital colleagues” or even run entire business processes independently.
“For the Employee Self-Service Agent, it was an easy and straightforward process. We made a presentation to people from Microsoft HR in France, which went well. So, when we went to the works council, there were not a lot of concerns. We got the green light very quickly.”
Isabela Cardoso, regional experience lead for France and Ireland, Microsoft Digital
One example of a digital agent that we’ve launched across the entire company (as well as externally to our customers) is the Employee Self-Service Agent. This AI-driven tool offers a “one-stop shop” that our employees can turn to for help with IT support, HR questions, and facilities requests.
Because the tool can access potentially sensitive personal information about employees, we were careful to make sure that our works councils were consulted during the internal deployment of this agent. Their experience reviewing Copilot—and the simplified process that Chemerys spearheaded—were key to winning rapid approval of the Employee Self-Service Agent.
“For the Employee Self-Service Agent, it was an easy and straightforward process,” says Isabela Cardoso, a regional experience lead for France and Ireland within Microsoft Digital. “We made a presentation to people from Microsoft HR in France, which went well. So, when we went to the works council, there were not a lot of concerns. We got the green light very quickly.”
Edith Dubuisson, a senior business program manager in Microsoft Digital who manages our relationship with the Microsoft works council in France, agreed.
“AI is a massive change, so getting the councils engaged early helps to deal with the fear and questions. We see the works councils as a true partner in this transformation of the company with AI.”
Edith Dubuisson, senior business program manager, Microsoft Digital
She stressed how important the concept of partnership is in making sure works council reviews go as smoothly as possible, especially with AI-related technologies.
“We make a point of including the works council in early discussions, telling them that we need them as a partner,” Dubuisson says. “AI is a massive change, so getting the councils engaged early helps to deal with the fear and questions. We see the works councils as a true partner in this transformation of the company with AI.”
Chemerys notes that as employee-created agents proliferate across the company, the review process will typically take place at the platform level, not the agent level. In this way, agents created with Microsoft 365 Copilot Studio will be treated much like tools created with something like the Microsoft Power Platform are handled.
“When it comes to low-code/no-code agents, you can compare the process to something like Power BI,” she says. “We’ve approved that platform to build reports, and then employees can create reports using it. In some countries, if an employee creates an AI agent using Copilot Studio that could impact the workplace in a sensitive way, then it’s their responsibility to get the proper approvals from our works councils. They can submit it through our standard process, which is why having that is so helpful.”
Embracing the future with our works councils
If there’s one true thing about technology in the age of AI, it’s that things continue to evolve at lightning speed. New tools and features are constantly being created, tested, and launched across Microsoft and many other cutting-edge, innovative companies.
Amidst all this rapid change, we continue to keep our works councils in the loop as these new technologies emerge. It’s a challenge that Microsoft is ready for, Chemerys says.
“There’s an avalanche of new solutions emerging all the time—so many different types of agents and other AI tools,” she says. “And the level is complexity is very high. But we have a great platform for works council reviews, so we can give them an early heads-up, which helps us maintain trust. They hate surprises, so we strive to stay ahead of things and make sure they stay informed.”
In the end, our works councils continue to be a source of invaluable feedback in this new fast-moving AI era. They play a role that transcends mere oversight and embraces proactive engagement, which makes for better products and happier employees—and customers.
Key takeaways
Here are some things to remember as you engage your own works councils with product reviews and discussions in the age of agentic AI:
- Engage your works councils early and often. Bringing them into conversations at the start—well before deployment—reduces uncertainty, surfaces valid concerns, and ensures smoother adoption of new AI tools like Copilot and employee-facing agents.
- Build trust through transparency and structure. A clear, well-documented approval process helps works councils understand new AI technologies and establishes the trust needed for productive, long-term collaboration.
- Simplify and unify communication channels. A single global request platform (like the one we use at Microsoft) improves coordination across works councils of different sizes, enabling smaller countries to benefit from shared expertise and creating a more consistent review experience.
- Balance innovation with worker protections. Structured tolerance phases, like those used for Microsoft 365 Copilot, allow employees to test new AI tools under controlled conditions while ensuring compliance with responsible AI principles and local regulations.
- Treat works councils as strategic partners in the agentic future. Their early feedback on digital agents—like our Employee Self-Service Agent—helps improve product design, accelerate approvals, and reduce fear or misconceptions about AI in the workplace.
- Design governance that scales with low-code and agentic tools. With AI agents proliferating, platform-level approvals—similar to the Power Platform model at Microsoft—ensure innovation can move quickly while still requiring review for individual high-impact scenarios.
- Stay ahead of rapid AI change with proactive communication. Works councils “hate surprises,” so providing early visibility into emerging tools helps maintain trust, reduces friction, and enables Microsoft to build better products for employees and customers alike.
Related links
- Learn how we’re deploying Microsoft Places with the help of our works councils.
- Find out how we’re becoming an AI-first Frontier Firm at Microsoft.
- Read the full guide on how we managed the internal deployment of Microsoft 365 Copilot at our company.
- See how we’re embracing the agentic future with our Customer Zero approach in Microsoft Digital.
We’d like to hear from you!
