Guiding hands: Inside the councils steering AI projects at Microsoft

As one of the first global enterprises to implement Microsoft 365 Copilot and other AI tools at scale, we entered the AI era with a mix of boldness and caution. That meant striving to capture the opportunities of AI fully while protecting our employees, customers, and company.

Throughout that process, we’ve relied on a series of employee-led councils to help us guide our strategy, drive our transformation, and shape our AI-forward culture. These bodies have been crucial for driving implementation excellence, maintaining an AI-ready data estate, and ensuring that we infuse this new technology with responsibility.

This is the story of how Microsoft Digital, the company’s IT organization, established the AI Center of Excellence (CoE), the Data Council, and the Responsible AI Council, and how these three groups are leading the way on effective, secure AI. Sharing these responsibilities across three different councils helps employees apply their specialized expertise and passion to challenges around implementation, data, and responsibility.

Different methods and structures will work for different organizations, but learning from the ways our AI councils operate can inspire your own efforts as you explore and implement AI projects.

The need for a guiding hand in AI

AI is a fast-moving technology. In just the last couple of years, the world has progressed through the advent of generative AI, the release of enterprise-grade AI solutions like Microsoft 365 Copilot, and the emergence of agents.

That velocity has meant that some companies have adopted AI with insufficient governance, security, data infrastructure, or strategies for aligning initiatives with their business priorities. That isn’t just risky. It also doesn’t drive effective impact.

“At Microsoft, we knew we couldn’t just implement AI for its own sake,” says Don Campbell, senior director of Employee Experience Success at Microsoft Digital. “Just like any other technology, the core challenge for AI is determining the right solutions to deliver on concrete, measurable business outcomes in the best, quickest, most responsible way.”

A clear vision detailing how AI will support your business is essential, but the technical and talent foundations also need to be in place. With technology as revolutionary as AI, that’s a substantial challenge.

Any organization adopting AI for the first time will have plenty of questions, and Microsoft was no different. In the past year, we’ve had to wrestle with some foundational ideas:

• What should our strategy be, and how do we get there?
• How do we enable employees through skilling and infuse AI into our culture?
• How can we organize our company’s data to support effective AI?

Answering those questions required insights across multiple disciplines.

Finally, the question of implementing AI responsibly was critical. Creating AI that’s safe, fair, and accessible to all ensures the AI revolution is a truly human movement.

A wide coalition came together from across Microsoft Digital and the company as a whole to contribute their expertise and tackle these challenges. Three AI councils have been instrumental in leading this work.

AI councils in action at Microsoft Digital

At Microsoft Digital, we have a long history of forming virtual teams that help us unlock agility.

Virtual teams are units composed of different disciplines and sponsored by senior leadership to help guide strategy for an area or initiative. Whether they’re mandated by the organization or arise organically through employees’ passion and initiative, they’ve become essential to the company’s fabric.

Not every organization will have the resources or staffing capacity to guide their AI initiatives in the same way as ours. But we’ve found that separate teams that reflect their members’ unique passions and skill sets help us zero in on specialized challenges across implementation, data, and responsibility.

The AI Center of Excellence

The AI CoE was Microsoft Digital’s first internal team dedicated to enabling AI. In fact, it formed before generative AI captured the world’s imagination.

At first, the team was simply a diverse group of professionals connecting over a shared passion for AI. But as our AI-powered transformation accelerated, they took on the responsibility for designing and championing how our organization uses this technology in the enterprise.

The AI CoE includes experts across Microsoft working in diverse fields. It’s a group of data scientists, machine learning engineers, business intelligence developers, research scientists, behavioral psychologists, and program managers.

This action-oriented team grounds its efforts in four workstreams underpinned by the company’s responsible AI principles.

As the maturity of AI-powered solutions has increased, the AI CoE’s role has also changed. At first, much of its work centered on ideation, education, and foundational architecture.

As AI initiatives have progressed, the team often showcases stories and learnings both internally and externally.

“Because we’ve connected our technical baselines and business priorities, we’re able to move faster,” says Faisal Nasir, principal architect with Microsoft Digital serving on the AI CoE’s leadership team. “Between strategy, technical architecture, implementation, and education, the AI CoE has been instrumental in finding the right direction for our community.”

One of the AI CoE’s current priorities is thinking through problems we’ve identified along our journey—issues like regulatory compliance issues or hallucinations. Combating these challenges relies on extensive collaboration with other teams, so their oversight and coordination are essential.

The AI CoE’s efforts also include retrospectives to help Microsoft Digital position itself for the successful execution of our AI roadmap’s next phase, helping to advance AI-powered transformation and adapt to new opportunities like agentic AI.

The Data Council

Our Data Council is a multidisciplinary team that includes professionals across several internal organizations, including Microsoft Digital, Human Resources, and Corporate, External, and Legal Affairs (CELA).

This team shapes a cohesive, scalable data strategy throughout Microsoft Digital that empowers cross-organizational insights and aligns with business goals. It was instrumental in implementing our data mesh architecture to enable domain-oriented data ownership, unlocking greater agility while maintaining security.

Much of the Data Council’s work is about resolving the challenges that massive quantities of enterprise data create. Of course, that mandate reaches beyond just AI, but good data is essential for AI to be trustworthy.

“Enterprise data management is a process that takes years, but AI moves fast, and organizations need to be efficient and enable teams with agility,” says Miguel Uribe, principal product manager lead responsible for coordinating Data Council workstreams in Microsoft Digital. “You can’t just live in the future, so you have to think about what to prioritize in the now to enable AI.”

This council addresses several data challenges:

• In an enterprise environment where there may be multiple copies of data, it’s important to determine which information is authoritative and surface that content through AI channels.
• Data drifts over time, and many people don’t yet have the knowledge or practices in place to keep AI systems in line with the most up-to-date information. Maintaining data freshness helps combat this drift.
• To surface the proper data, the data needs to be discoverable. In an enterprise environment like Microsoft, with no less than 19 enterprise data lakes, that can be a challenge.
• Effective governance ensures ownership over data, determines whether AI should be able to access it, and what policies or regulations apply to different kinds of data. Establishing proper governance requires extensive planning, implementation, and ongoing work.

“At the heart of our enterprise Data Council strategy is a bold commitment to unifying people, process, and technology to power AI with purpose,” Nasir says. “By aligning governance, data quality, and culture, we’re not just managing data—we’re also turning it into a trusted foundation for intelligent insights, ethical innovation, and enterprise-wide transformation.”

Our vision for our data council is to build and employ a cohesive data strategy across Microsoft Digital that enables descriptive, predictive, and prescriptive analytics while delivering intelligent insights through connected and scalable data capabilities and platforms. Our team tackles these challenges by prioritizing the five following strategic pillars that work across the four workstreams that follow.

The Data Council’s most recent focus is accelerating our AI journey at Microsoft Digital through “AI-ready data. That means secure, compliant, and reliable enterprise data we can use to enable the AI applications and services we create to boost our enterprise processes.

“Thanks to AI, we’re in the era where data is a part of every role,” says Diego Baccino, principal software engineering manager leading the Technology Strategy workstream for Microsoft Digital. “The world isn’t siloed into ‘data analysts’ and ‘everyone else’ anymore, so it’s our job to present data in a way that works for everybody.”

The Responsible AI Council

The Responsible AI Council is part of a much larger fabric that reaches into every organization across Microsoft.

Responsible AI is a top priority at the company, and it requires both internal policymakers to develop its underlying principles and operational leaders to infuse those principles into all the work we do. With a discrete council covering responsible AI, we can provide better oversight for AI projects and support for the employees creating them.

Microsoft established the Office of Responsible AI in 2019. It determined that every AI initiative at the company should receive an impact assessment to make sure it’s properly envisioned, scoped, and secured. These reviews ensure our internal AI projects align with the Microsoft Responsible AI Standard, our rulebook for people developing or using AI.

At the same time, members of different teams across Microsoft were already forming communities of practice around responsible AI. Organic virtual teams within Microsoft Digital and other business groups with a natural interest in technology, law, and governance were already exploring ways to use AI responsibly. With the Office of Responsible AI’s directive around impact assessments, these volunteer virtual teams became reviewers for AI initiatives within their respective organizations.

Over time, we formalized this process into the One Responsible AI (OneRAI) tool. This portal logs every project. It pulls in Responsible AI Council champions to review AI projects and provide direction or education where necessary. And crucially, it guides AI developers through initial impact assessments and final release assessments.

As members of the company’s IT organization, Microsoft Digital’s Responsible AI Council was instrumental in helping create this tool. It’s now in operation across the company and helps instill responsible AI standards into the software development lifecycle behind AI projects.

“We didn’t want this to be just a compliance step,” says Jamian Smith, principal product manager and Responsible AI co-lead at Microsoft Digital. “We wanted it to be an opportunity for reflection and guidance to really incorporate responsible AI thinking into the product lifecycle.”

As representatives of the Office of Responsible AI within Microsoft Digital, our Responsible AI Council carries the torch for six core principles and includes champions across several key areas of our work.

Aside from managing the OneRAI tool, Microsoft Digital’s Responsible AI Council meets regularly with the Office of Responsible AI to keep both groups aligned as our transformation progresses. They also release quarterly newsletters to reinforce the impact assessment process and highlight changes to responsible AI policy, all while connecting with the teams creating AI projects to provide guidance and check progress.

Aligning our efforts with Microsoft’s vision for AI

As our AI councils have grown and matured, we’ve looked for ways to align their work even more closely with Microsoft Digital’s vision. That vision depends on leadership.

To help coordinate the councils’ efforts horizontally, we created a set of pillars and a corresponding team of leaders to own accountability for those areas. That team aligned on three core priorities:

• Transform and secure our network and infrastructure to infuse operations with data-driven insights, improve reliability, and revolutionize productivity and compliance by preemptively addressing potential disruptions.
• Revolutionize user services across tenant management, collaboration services, and support, with the goal of making Microsoft employees the most connected, efficient, and productive in the industry.
• Accelerate corporate functions growth by streamlining processes, ensuring compliance, and transforming actionable insights into real-world outcomes to equip function-specific roles with the tools and insights they need to excel.

A complementary team of executive sponsors and leads is accountable for supporting each of these pillars horizontally across the company. Together, these roles shape and influence the vision, strategy, and execution of our AI-powered digital experiences. While pillar sponsors provide support and direction to align outcomes with Microsoft Digital’s overall vision, pillar leads are accountable for driving the work forward and ensuring it matches our business goals.

As the councils’ roles have evolved, there’s been greater collaboration between them. Their collective efforts make for rich, organic cooperation and strong connections across priorities. Both the AI CoE and Data Council meet regularly with the pillar leads as the Responsible AI Councils work in parallel. All of this provides a way to synchronize different workstreams with Microsoft’s vision for AI.

“The closer we work together, the more effective we are at addressing our priorities,” Uribe says. “At the end of the day, it’s all about delivering on our strategy with agility by continuously adjusting our plans to what’s most needed right now.”

Microsoft’s scale and sophistication make discrete but collaborative councils necessary. If you’re operating on a leaner basis, the exact structure isn’t important. What matters is prioritizing strategic implementation, AI-ready data, and responsibility, then putting the right people in place as custodians of those crucial components.

Our experience can act as a guide as you consider how to organize your own AI guiding hands.

Empowering AI initiatives to drive business impact

As each of these teams works within their purview to guide Microsoft Digital’s AI initiatives, they’re committed to continuous improvement.

“Organizational teams naturally operate at varying speeds of innovation and adoption,” says Nitul Pancholi, principal and director of the AI Center of Excellence. “While each group makes meaningful, incremental progress, the true value lies in creating forums where we can openly share our diverse experiences and insights. Enabling this cross-pollination of ideas accelerates collective learning and drives strategic alignment across the enterprise.”

The systems that the AI CoE, Data Council, and Responsible AI Council have established are weaving vision, AI readiness, and high standards of behavior into every project we undertake. That means we’re progressing steadily toward an AI-powered future, complete with robust infrastructure and solutions that meet the demands of a rapidly changing enterprise.

One of the most important developments is the introduction of a framework for tracking our AI initiatives’ value and impact. It’s a way for us to identify our big bets and evaluate them across six dimensions of impact:

• Revenue impact: Direct contributions to revenue generation and business growth
• Productivity and efficiency: Efficiency gains while completing tasks and processes without a reduction in quality
• Security and risk management: Improvements in identifying, preventing, and managing security vulnerabilities and risks
• Employee and customer experience: The impact of AI initiatives on employee satisfaction, engagement, and productivity
• Quality improvement: Enhancements in the quality of deliverables, services, and processes
• Cost savings: Reduction in operational costs and resource allocation efficiencies

From there, we zero in on specific metrics to measure their value. The goal is to attune our AI projects with Microsoft’s overarching business objectives so they drive the greatest impact.

This measurement framework is laying the foundation for the insights necessary to empower a continuous improvement mindset. As a result, we’re looking ahead to a virtuous cycle of implementation, insights, and improvement for all AI initiatives.

All of this depends on the diligent work that our council members perform day in and day out. As we progress into greater AI maturity and move from enablement to impact, the groundwork these teams have laid will be the foundation for our continued innovation in the AI era.