Visma continuously investigates how artificial intelligence is used within business software, balancing central coordination with local autonomy. At the heart of this transformation is a dedicated AI team that drives adoption internally, enabling product innovation and supporting the company's diverse portfolio in Europe and Latin America. From chatbots, generation AI, and agent AI: Business software progress is fast.
As AI matures, Visma moves steadily from experiments to scale. According to Jacob Nyman, AI Director at Visma, the key is not to direct strict strategies from the top, but to create a framework where innovation can flourish. “It's really not wise to pretend to know exactly how to do things from a central perspective,” he emphasizes.
Visma's AI conversion revolves around four interconnected shifts. The first is to build an AI-native workforce that embeds AI into daily work to amplify skills and productivity. The second is AI-Native product development that uses AI throughout the software development lifecycle and accelerates delivery and improves quality, from prototyping to coding support. The third is the creation of AI-Native products, providing intelligent features directly to the customer. Finally, AI-Native Growth features leverage AI to increase customer operations such as support, sales and marketing, driving both efficiency and a stronger customer experience.
This structure ensures that AI is not only used internally, but also embedded in the product and tailored to sector-specific expertise. Nyman emphasizes that each layer strengthens the other layers. Internal skills provide confidence to innovate your product, and successful customer implementation encourages employees to adopt more AI.
Developer first adoption as a foundation
Visma's most successful adoption case to date is one of the developer community. The company initially saw the use of 5-7% of tools like Github Copilot. Today, almost every technology employee uses AI-assisted coding every day. Developers also experiment with alternatives such as cursors and windsurfs, demonstrating that diversity of tools can help accelerate adoption.
The logic is simple. When developers build using AI, they build a lifecycle from design to delivery benefits. This is not limited to writing code. Customer teams also rely on AI to improve service delivery and are an area where customer support is outstanding. In parallel, employees across business functions use productivity and specialized tasks using Google's workspace ecosystem, including Gemini and AI Studio.
By embedding AI in both the technology and operational layers, Visma ensures that AI is not just an isolated experiment, but a horizontal ability to support the entire organization.
Hundreds of big AI front row seats
One of Visma's unique challenges is managing AI across the enterprise. Globally, over 100 software companies are part of the group. Each operates in a market with a variety of regulatory requirements, customer expectations and cultural contexts. Centralization can suppress innovation, but everything can leave a completely localized risk fragmentation.
The solution is in a central enablement along with local decisions. AI-Team guarantees access to APIs, tools and vendor partnerships, as well as security and compliance guidance. However, the responsibility remains with each company to decide how to use AI. “We are a group that is coming together and we can partner with some very good global leaders within AI,” explains Nyman. “However, we do not want to override local decisions. The presence of our region in a regulated market is a competitive advantage.”
This coalition approach ensures that AI adoption meets local business contexts while maintaining trust and compliance standards across the group.
– The text follows the box below –
Enablement, Acceleration, Optimization
Visma assembles its transformation in three different phases.
- Enabling – Lay the foundation by ensuring access to tools and powerful vendor partnerships in favorable terms.
- Accelerate – Acquire successful practices from individual companies and distribute them across the group to speed up adoption.
- Optimization – Resist the temptation of early optimization in such rapidly moving areas, whilst integrating frameworks and technologies that have proven to be the most effective.
This step-by-step approach ensures steady progress while leaving room for adaptation. It also demonstrates Visma's perception that transformation is not a one-off project, but a continuous cycle of learning and scaling.
Next frontier
One of the most powerful signals of AI maturity is the transition from simple assistants to autonomous agents. If a chatbot once answered a question, the agent now handles workflows across multiple systems. Nyman describes this as a natural progression. “From chatbots to copilots, assistants and agents.” Agents became the central theme across enterprise technology in 2024. Companies want automation that allows them to run the entire process beyond the support query. But, as Nyman points out, this shift isn't just about branding existing tools differently. “We don't want to pretend we have agents. We want to create agents,” he says.
The difference between hype and reality lies in engineering. A true agent requires an infrastructure that allows dynamic interaction with multiple systems. This is unscrupulous to manage it with traditional APIs and integrations alone. Therefore, new protocols like the Model Context Protocol (MCP) are attracting attention as they allow for more sophisticated context engineering.
Reliability issues composite engineering assignments. Giving agents autonomy introduces new risks. “When they have the freedom to do more, they can also do more wrong,” Nyman admits. Even when agents automate routine tasks, human surveillance remains crucial for important decisions. This hybrid balance of autonomy and reliability defines the most effective developments of today.
Agents stay here
Visma's experience confirms that successful AI agents are based on deep domain knowledge rather than generic intelligence. In Norway, for example, lawyers worked closely with AI developers to ensure that tax and accounting agents reflected accurate legal expertise. In France, the teams incorporated regionally-specific regulatory knowledge directly into their applications.
This principle extends to verticals in all industries. The closer AI agents are tied to specialized knowledge, the more effective it becomes. Generic agents may handle surface-level tasks, but the real differentiation is encoding the expertise already owned by the company.
Despite current limitations, agents represent more than passing trends. Several factors promote their persistence. Geopolitical competition ensures that states invest heavily in AI. The human brain itself proves that efficient and general intelligence is possible. And most importantly, agents provide intuitive abstractions to businesses. You can organize, train and manage like an employee.
“Agents attack this complete level of abstraction, limited enough to design, but capable enough to do amazing things,” recalls Nyman. This framing resonates with business users, accelerates recruitment and bridges the gap between technological innovation and everyday workflows.
A practical vision of AI in software
Visma's AI Journey shows that corporate recruitment is not about following the latest trends, but about building trust through activation, experimentation and careful scaling. From developer-first adoption to open source platforms, regional autonomy to central partnerships, every step reflects a practical balance of innovation and reliability.
As agents move from hype to production reality, companies combining technical engineering and domain knowledge separate themselves from noise. For Visma, the long-term vision is clear. AI is not just a feature that is fixed to the product, but also a feature that is woven into the structure of the company. It was centralized in places where it made sense and localized to places of importance.
Tip: Domain-specific AI beats popular models in business applications
