A lesson for leaders: The deliberate AI risk we took
When we at Sofigate moved into AI, we did it boldly and gave our people free rein to experiment. This made the tools familiar, but did not show as a positive result in the bottom line. Katri Kolesnik and Essi Hirvonen explain why the deliberate risk still paid off – and how companies can build their people’s AI capabilities and business value at the same time.
When ChatGPT and other generative AI tools began entering working life, everyone at Sofigate was immediately given a licence for an AI tool of their choice. The same happened when it became possible to build agents: we encouraged everyone to create agents to support their work.
The idea was simple: use them, try them, learn them.
We knew that, from a business perspective, the early results would be modest – and they were. Costs increased, internal expenses began to rise, and no measurable financial benefit followed.
Business Executive Katri Kolesnik says Sofigate now advises its customers to clarify their goals first and only then turn to the tools. Still, our calculated risk proved fruitful, especially from a learning perspective.
“As a business technology company, it was important for us to move boldly into AI. That made AI familiar to our people, and many enthusiastic experimenters became important pathfinders for the rest of the organisation. The journey also gave us lessons that now help us help us support companies in developing their people’s AI capabilities and connecting adoption more quickly to business value,” Kolesnik says.
Not everything should be taught to everyone
Kolesnik is responsible for people capability development and the training business at Sofigate. She says AI capability needs to be divided into two things: a common foundation for everyone, and deeper expertise tied to the organisation’s own goals.
The first applies to every employee, from coders to CEOs.
“This is about AI literacy. Everyone needs to understand what AI is, what it is worth using for, where its limits are and what risks are involved in using it,” Kolesnik says.
But not everything should be taught to everyone, and this is where many companies go wrong.
“If an organisation has not first defined what it plans to use AI for, training can easily go to waste. A person may learn to build an agent, only to realise shortly afterwards that it supports an outdated way of working. In that case, it can add to the workload instead of freeing up time for more important tasks.”
The same phenomenon is visible more widely in working life. According to our latest playbook, as many as 73 per cent of organisations are still automating existing tasks instead of using AI to renew work and processes.
“That is exactly why training cannot start with the tool. It has to start with what the company intends to change in its work. Only after that can we ask what kinds of tools and support are needed in different roles.”
Licences and training days are not enough
Sofigate’s Essi Hirvonen, who works as a project manager in a client organisation, remembers how her own AI learning path began with a ChatGPT licence in winter 2023.
“We were actively encouraged to use AI. We were even so quick that there was not yet a very clear model for what AI should be used for,” Essi laughs.
Soon, training also became available. For Hirvonen, it provided a much-needed foundation: an understanding of how AI works, ethical ground rules, source criticism and prompt writing.
“You do not learn AI through theory, but by doing. A one-day training session often sparks interest and introduces you to the subject, but if the new skills are not needed and supported in everyday work, they are forgotten. That is why, for me, easily accessible support for learning new AI skills has been at least as important.”
For Hirvonen, Sofigate’s AI agents, which know the company’s business models and knowledge work inside out, have been the best way to support her own learning. They help in client work, where new situations arise constantly and there is not always time to ask a human colleague.
According to Kolesnik, AI is changing the very logic of learning. Skills are no longer built only through training sessions or online courses, but through continuous interaction with AI alongside everyday work.
“With AI tools, learning is more personalised. Help is available exactly when the need arises,” Kolesnik says.
Capability costs money
Kolesnik’s most important advice is simple: developing capabilities costs money.
“We may even have paid a little too much for those early lessons, but on the other hand, our
Sofigate invested three million euros in employees’ AI capabilities, AI agents, new ways of working, and service development.
“That is not a small sum, around 4,000 euros per employee.”
According to Kolesnik, this is exactly the question every management team should put on the table. Are we prepared to pay for our people’s AI capabilities, and how much?
“Training has traditionally been cut whenever the economy tightens. In the age of AI, that logic is dangerous if a company genuinely wants to renew itself through AI,” she says.
If you do not train your people, someone else will
AI capability has also become a factor in attracting and retaining employees. People understand that AI skills affect their own value in the labour market.
“If the employer does not invest in capability, many people will develop it in their free time. Or they will go somewhere where investment is made in learning,” Hirvonen says.
Hirvonen is pleased that Sofigate moved into the use of AI boldly. She says AI should not be seen as a threat that takes work away from people, removes the need for human thinking or narrows opportunities for younger employees.
“Our experience is that it can do the opposite: help people learn more independently, encourage them to take on more demanding tasks and teach them to use their own professional expertise to evaluate the solutions produced by AI,” Hirvonen says.
Kolesnik also says courage and open-mindedness are strengths in AI adoption, and they were strengths at Sofigate too.
“The most important thing for companies to understand is that licences and training only open the first door. People also need a reason, the skills and the support to step through it.”
Want to read more about leading work in the age of AI?
Download our new playbook, There is no more work without AI.
The experts:
Katri Kolesnik is responsible for capability development in Sofigate’s Business Renewal team. She has built and led Sofigate’s training business and has also worked for many years with the Business Technology Forum. Kolesnik is especially interested in how people and companies can renew themselves through AI.
Essi Hirvonen works as an Advisor at Sofigate. In her client work, she focuses primarily on change management and works as an IT project manager. Hirvonen does not have a coding background, but she is all the more enthusiastic about learning to use AI. Today, she makes full use of AI to support her everyday work.
Leader, don’t start with AI – start with the problem
Your next team member might not be human. Are your managers ready?
