Technology vs Behaviour: What the Carlsberg Copilot journey taught us

Organisations often assume that once the technology is in place, the change will follow. 

A new platform is launched. A new tool is rolled out. And the system works exactly as designed. Yet the impact takes longer than expected. 

Because technology doesn’t drive change. Behaviour does. 

In today’s digital workplace, organisations are investing heavily in platforms like Microsoft 365 and tools such as Copilot. These technologies have enormous potential to improve productivity, collaboration and the way teams work together. But potential alone doesn’t create transformation. 

For real change to happen, people need to understand why the technology matters, how it fits into their work, and what new behaviours it enables. Without that clarity, even powerful tools can remain underused. This is where many transformation efforts run into friction. 

The technical implementation succeeds, but behaviour takes longer to shift. Teams fall back on familiar habits. New features go unexplored. The opportunity is there. It just hasn’t fully translated into everyday work. 

That is exactly the gap we helped Carlsberg bridge in their Copilot adoption project. 

How Open supported Carlsberg: from rollout to real habits 

At Open, we focus on the people side of organisational change. In this project, our role was never “the technical rollout of Copilot per se”. It was the human rollout. The communication, the engagement, the learning journeys, and the behaviour change that makes adoption last. 

We worked closely with Carlsberg’s M365 service owner and teams across the business to move from a small pilot to organisation-wide adoption. The project scaled from roughly 500 Copilot licenses to more than 25,000, but the real story is not the number. It is how Carlsberg made Copilot part of the way people work. 

From the kickoff workshop, we aligned on two parallel tracks: 

• A people-first track: how Copilot would become everyday habits across markets and functions 

• A business track: how Carlsberg would measure adoption and value, so senior leaders could follow progress and make confident decisions 

From there, we helped shape a structured plan and brought classic change levers into an AI context: a strong narrative around the purpose of Copilot, communities and champions, function-based training, and practical formats that helped people start small and build confidence over time. 

Five takeaways from Carlsberg’s Copilot adoption (with the “how” behind them) 

1) Copilot adoption was a behaviour change initiative, not an IT rollout 

Carlsberg treated Copilot as more than a technical deployment. The move from 500 to 25,000+ licenses only worked because the focus was on how people work, not just what tools they had access to. 

Our contribution was to design adoption around habit-building and practical use cases. Not feature training for its own sake, but confidence, repetition, and “tomorrow-morning” value. That is how behaviour shifts across markets and functions. 

Carlsberg learning: Scaling Copilot required changing behaviours, not running a classic IT pilot. 

2) “AI out of IT” unlocked real ownership across functions 

Carlsberg kept platform ownership in IT (as it should be), but deliberately moved learning and usage ownership into the business. HR (PNC), Sales, Marketing and other functions were encouraged to define what Copilot should do for their work. 

We supported this by creating formats where functions could articulate their needs and translate them into use cases. We also helped build the “permission structure” that made it acceptable for functions to own learning agendas, rather than waiting for central delivery. 

Carlsberg learning: Momentum came when functions felt it was their tool, not something delivered by IT. 

3) Trust was built through honesty, not hype 

A striking feature of the project was the commitment to transparency. In communication and training, limitations were acknowledged rather than oversold. This reduced skepticism and increased long-term adoption. 

We helped reinforce this approach in messaging and facilitation. When people feel the organisation is honest about what AI can and cannot do, they are more willing to experiment without fear of being misled or judged. 

Carlsberg learning: Sustainable usage came from realistic expectations, not silver-bullet storytelling. 

4) Agents scaled because employees built them themselves 

Carlsberg ended up with more than 2,500 agents, and around 85% were built by employees. Most were small, local and function-specific. Some were short-lived, and that was not a failure. It was learning in motion. 

We supported this through both broad and targeted initiatives. One example was an innovation challenge that invited employees to propose ideas and build simple one-pagers. Another was more structured function processes: inspiration sessions (including “tap room sessions”), mapping a workday, identifying friction points, and developing recommendations for which agents could help. In some cases we also supported value assessment (time saved, quality improved) and technical validation with solution architecture, so functions had a clear path: “can I build this myself, or do I need help?” 

Carlsberg learning: The goal wasn’t perfect agents. It was empowering experimentation that solves real problems. 

5) Copilot became a driver of inclusion and cross-market collaboration 

One of the most meaningful outcomes was cultural. Copilot supported accessibility and inclusion, including employees with dyslexia and colleagues with deep-thinking preferences who benefit from time to process and articulate ideas. It also strengthened collaboration across Carlsberg’s many markets and functions by creating shared ways of working and shared learning. 

We helped nurture the community and communication layer that makes those effects spread. Adoption becomes cultural when people share use cases, teach each other, and see leaders role-model the behaviours. 

Carlsberg learning: One of Copilot’s strongest effects was enabling more people to contribute, across roles and geographies. 

The conclusion: technology creates opportunity, people create change 

At Carlsberg, Copilot became part of everyday work when the organisation stopped treating it as a tool rollout and started treating it as a shared change in how people work. 

That is the heart of digital transformation. When you invest in both the platform and the people side, change does not just launch. It anchors.