F-Secure · April 2025–March 2026
One year.
From AI tools to leaders building agents.
We started with tools and experiments in engineering. Twelve months later, leaders were building agents for their own work and the change had spread into business functions.
Antti Tevanlinna — my year at F-Secure.
I worked inside F-Secure five days a week for a year. I started in engineering in April 2025. By January 2026, I was working on the change across the whole company.
I created the engineering and business training tracks with the team, acted as head trainer, trained the leadership personally, and developed internal trainers. I also built context for agents to use, ran workshops, and wrote 30 transformation newsletters.
The work grew through colleagues who built, demonstrated, taught, and carried it into their own functions. My job was to make that possible and keep it moving.
How the year unfolded
April–June 2025
Get tools into people’s hands.
We chose Claude as the standard tool and made signup easy. A company-wide hackathon started the experiments. I formed an ambassador network, started weekly demos, and launched a newsletter that named people and linked their work. The call to action was simple: try AI and share.
August–September 2025
Connect agents to the actual work.
Integrations gave agents access to internal systems. I started building curated, version-controlled context about products and architecture. Agents could work from company knowledge. A product team working with agents from the outset gave others a concrete example to follow.
October–December 2025
Build competence. Develop the trainers.
We launched hands-on engineering training on real work. Engineering metrics helped us look beyond enthusiasm: improvements varied across teams, and some showed no change.
At the same time, the leadership team was shaping the company’s AI vision. It grew out of their own thinking and experience.
January–March 2026
Get leaders building personally.
I launched the business training. The leadership team enrolled themselves. The CRO built dashboards and demonstrated them at a company townhall. Roughly 40 business people were building agents for their own work, across functions including product, sales, support, and leadership.
What began in engineering was becoming a company-wide change. Leaders could act on possibilities they had experienced themselves.
What carried the change
Hands-on training built competence. Demos made colleagues’ progress visible. The programme developed 15 internal trainers and reached more than 200 engineers through the train-the-trainer model. The weekly newsletter kept examples, people, and lessons in circulation.
Leaders building personally changed the conversation. They could make more concrete decisions about where agents belonged because they had done the work themselves.
Where it was harder
Knowledge sharing did not scale through encouragement alone. Cross-functional workshops produced useful findings. But when the follow-up work wasn’t prioritised, it stalled. Access to tools and a good workshop were only the beginning.
If I started again, I would bring team leads and directors into hands-on training earlier. They shape daily working practices. I would also invest sooner in curated context, reliable checks, and the way teams learn from each run.
What I learned
The sequence that mattered was people building, discovering what was possible, and carrying that knowledge into bigger decisions. The training and the work around it gave the organisation a way to keep learning.
That experience informs Agents 101 and Agentic Engineering 101 today. Both have continued to evolve since this work.
Get your people building.
Tell me what you want your people to be able to do. We’ll work out where to start.
Talk to Antti