How Visma Accelerated AI Adoption Across 5,000 Developers with GitHub Copilot

AI adoption starts with leadership

When generative AI began reshaping the software industry in 2024, Visma – one of Europe’s leading companies in the field – decided not to wait on the sidelines. With Eficode’s support, this approach led to a company-wide transformation in how software is built at Visma.

Visma’s AI journey started with a strong mandate from the management.

Building a foundation for AI-assisted development

It soon became clear that successful AI adoption required more than access to tools. Visma launched an internal training program on GitHub Copilot and delivered intensive two-day workshops across multiple teams and business units, reaching altogether over 2 200 employees.

This built a strong foundation for AI-assisted development. But as AI tooling advanced at accelerated rate, Visma needed to continuously level up their capabilities. That is where Eficode stepped in.

The companies designed a large-scale training initiative focused on highly advanced Copilot usage, AI workflows, and hands-on adoption. The program included 30 advanced three-hour workshops, with 25 participants per session. In addition, deep-dive webinars were available for Visma's 5,000 developers.

Raising the floor, not just the ceiling

The training program started in February and progressed rapidly. The main objective was not so much to maximize the top-notch expertise but rather elevate the broader group.
"The top five percent will always figure things out themselves. The real impact comes from raising the floor and making sure that the middle ground comes on board", Alanen says.

This philosophy became central to the collaboration. Some of Visma's developers were already operating at an extremely advanced level, with certain teams generating up to 95% of their code with AI. Others remained hesitant to engage with AI-assisted workflows.
"This is not a question of skills anymore; it's a shift in mentality. AI is the biggest thing in software engineering in the last 40–50 years, and people need time to adapt. For some, AI initially felt like it would take away what made the work enjoyable. But there is so much fun to have with AI! Many are now discovering that AI allows them to focus on bigger and more creative engineering problems", Alanen says.

From AI features to AI workflows

One of the major learning points during the project was the transition from AI features to workflows.

"In the beginning, everybody focused on features. Now the real question is how to turn them into scalable AI-driven workflows that improve software engineering", Müller says.

Visma sees the next major transformation in autonomous AI agents and cloud-based delegation.

"I want to see full delegation: AI agents working independently in the cloud while engineers focus on direction, architecture, and outcomes. I've started to advocate for 'No manual coding week'."

For Visma, this evolution is already reshaping the definition of software engineering.

"Manual coding is dead, long live software engineering", Alanen states.

Measuring AI adoption beyond productivity metrics

Visma continues to explore how to measure AI adoption and productivity gains. The company uses surveys, before-and-after assessments, and internal Copilot usage trends to monitor progress. However, meaningful measurement remains challenging.

"GDPR, ethics, and practical limitations all make this difficult," Alanen explains. "Once a metric becomes the goal, people start optimizing for the metric instead of the actual outcome."

Instead of focusing purely on quantitative productivity numbers, Visma concentrates on long-term business outcomes and developer enablement. The economic rationale for AI is still very clear.

"If a developer costs €10,000–15,000 per month and an AI tool a few hundred euros while making them significantly more productive, it's a no-brainer."

Enabling people for long-term AI transformation

For Visma, the collaboration with Eficode demonstrates what is possible when leadership commitment, internal enablement, and practical AI training align.

From Eficode's point of view, Visma is a role model in their field. They invested early, trained at scale and created a culture where developers are encouraged to experiment and adopt AI in practice.

"We see where the customers are and it is our mission to lift them even further", Müller says.

Visma has understood an important reality of AI transformation: it is all about enabling people.

"It's not happening in the brain, it's happening in the heart. The information is out there – you just need to take the mental leap", Alanen concludes.