Femke Cornelissen on AI at Carlsberg: from technology to organizational transformation

Femke Cornelissen, Chief Transformation Officer at Wortell, regularly “interns” at different organizations to experience firsthand how AI is applied in practice. In this series, she shares her insights. This time: Carlsberg, where AI is not seen as a technology project, but as a fundamental organizational transformation. During her visit, she spent a day shadowing Oliver Munk Winther, Microsoft Copilot Lead & Service Owner M365 at Carlsberg Group, to experience from the inside how this transformation is taking shape.

It quickly became clear how deeply AI is embedded in the way of working. But even more striking was how strongly culture, leadership, and collaboration shape its success.

Trend 1: Culture as the foundation for AI success

At Carlsberg, everything starts with culture. New employees are actively immersed in the company’s history, as understanding that history is considered essential to truly becoming part of the organization.

This sense of pride is not abstract. It is visible and tangible in everyday work. Employees wear Carlsberg pins, speak about the brand with conviction, and see themselves as part of something bigger.

A strong culture fosters engagement, encourages knowledge sharing, and creates openness to change. As a result, AI is not experienced as something “new” that needs to be learned, but as a natural next step in how work evolves. The willingness to learn from one another is already embedded in the organization’s DNA.

Lesson learned: without a culture centered on learning, sharing, and ownership, AI remains an isolated experiment. With the right culture, AI becomes part of everyday work.

Trend 2: AI is not an IT project, but an organizational challenge

One of Carlsberg’s most defining choices is deliberately positioning AI outside of IT. While IT naturally plays a role, responsibility for adoption and usage lies with the business and departments such as People & Culture.

This has a major impact. Employees do not see AI as “just another IT tool,” but as a natural way to work smarter.

It also ensures that adoption is driven not by technology, but by real organizational needs. Stakeholders outside IT, such as communications and HR, play a crucial role in shaping the narrative and actively engaging employees.

With over 2,500 Copilot licenses and an adoption rate of around 97%, AI is deeply embedded. But more importantly, success is not measured in licenses. It is measured in behavioral change.

Trend 3: Leadership makes the difference 

Leadership plays a key role in AI adoption at Carlsberg. Executives actively engage in discussions about the impact of AI on decision-making, roles, and responsibilities. This shifts the conversation from “what can the tool do?” to “what does this mean for how we work?”

In the early stages, adoption was primarily bottom-up, with employees exploring applications on their own. Today, this is complemented by a strong top-down component, including one-on-one sessions with senior leaders to actively involve them.

This combination ensures that AI does not exist alongside the organization, but becomes part of the core of how work is done.

Trend 4: From possibilities to choices 

Although Carlsberg has access to hundreds of AI agents and a broad Copilot rollout, the focus is not on maximizing functionality, but on making deliberate choices.

AI is applied where it truly adds value, such as:

• Demand forecasting and supply chain optimization
• Contract analysis and risk detection
• Support for marketing campaigns and content creation

In the early stages, collecting use cases proved difficult and often unsustainable. They quickly became outdated or were not used. As a result, the focus has shifted to enablement: teaching people how to apply AI, rather than handing them ready-made use cases.

Lesson learned: Mature AI is not about doing everything. It’s about knowing where you create impact.

Trend 5: Adoption is a prerequisite, not the end goal

Carlsberg’s high adoption rate is no coincidence. It is the result of a deliberate approach. But the focus is not on usage; it is on change.

The real questions are:

• Are employees making better decisions?
• Is collaboration changing?
• Are meetings becoming more effective?

In practice, tools like Copilot make meetings more efficient through summaries, action tracking, and preparation. But the real value lies in how people start working differently as a result.

Adoption is supported by a combination of communities, a central adoption hub, and a network of champions. However, the greatest impact comes from hands-on sessions and direct application in daily work.

Trend 6: Learning works differently than you think 

While many organizations invest in extensive training programs, Carlsberg deliberately takes a different approach. Traditional feature-based training proves to be of limited effectiveness. Instead, the focus is on short, practical sessions that show how AI actually changes the way people work.

Initial sessions provide a foundation and create a safe space for questions. After that, teams take ownership and develop their own ways of working with AI. Notably, sessions focused on “how work changes” attract significantly more participants than those focused purely on features.

Trend 7: Data as the foundation

AI is only as good as the data it runs on. That’s why Carlsberg has invested heavily in data quality and governance in recent years.

Particularly within SharePoint, where most data resides, processes have been established for cleanup, version control, and structure.

This ongoing focus on data quality ensures:

• More reliable outputs
• Greater trust in AI
• More consistent processes

Data is therefore not just a prerequisite, but a key enabler for scalable AI.

Trend 8: Agents: innovation first, governance later

A notable choice at Carlsberg is how they approach AI agents. Instead of implementing strict governance upfront, employees are given the freedom to experiment.

The idea is simple: innovation does not come from control, but from ownership.

By starting early and minimizing barriers, use cases emerge organically from the business rather than being driven solely by IT. Governance then evolves alongside usage.

"Don't shoot after the tree in the backyard. Shoot for the stars."

An approach that encourages experimentation and accelerates innovation.

Trend 9: Impact on collaboration and inclusivity 

At Carlsberg, AI impacts not only productivity, but also collaboration and inclusivity. Meetings are shorter, better prepared, and more consistently followed up. But perhaps more importantly, AI helps employees participate more effectively in conversations.

In a pilot group in the UK, employees with dyslexia or those on the autism spectrum were able to participate more actively in meetings and articulate their ideas more clearly. AI thus functions not only as an efficiency tool, but as an amplifier of human capability.

AI is organizational change 

Carlsberg demonstrates that successful AI adoption does not start with technology, but with the organization itself. It requires:

• Positioning outside of IT
• Active leadership and dialogue
• Strong data foundations
• Space for experimentation and ownership
• A focus on behavior rather than tooling

AI is not an IT project. It is a fundamental shift in how people work, collaborate, and create value.