"On Our Own, We Might Have Done Something Roughly Similar" – How FAIR Helped ReactorEd Get It Right from the Start
ReactorEd, the company behind a language-learning platform for teachers, was already familiar with AI tools when it began exploring new opportunities with support from FAIR. The result of the collaboration was a new AI tutor, and expert guidance helped the company accelerate development while avoiding common pitfalls.
Text and photo by Eemeli Sarka, 8.6.2026
Before working with Finnish AI Region (FAIR), ReactorEd had already automated parts of its exercise creation process using AI. With its existing task generator, teachers could automatically create sentence and vocabulary translation exercises from any text, and the company’s ChatGPT integration enabled multiple-choice questions and vocabulary activities. However, it quickly became clear inside the company that there was potential to do much more with AI.
The collaboration with FAIR began at a pop-up event, where ReactorEd’s digital maturity and AI readiness were assessed and future needs and opportunities were discussed.
– In the early stages we brainstormed ideas with FAIR and evaluated different directions. We eventually decided to develop a Proof of Concept for an AI tutor that could automate dialogue exercises in language learning, says ReactorEd CEO Kari Savolainen.
The goal was to build a tool that allows teachers to create AI tutors students can converse with. Conversations could happen in different roles and situations depending on the learning objective.
– Unlike predefined dialogue scripts, conversations with the AI tutor can continue endlessly. It can, for example, simulate a work-related customer service situation, Savolainen explains.
Straight to the Point
According to Savolainen, development got off to a fast start. Together with FAIR experts, ReactorEd defined clear objectives for the PoC, agreed on the number of expert hours available, and set a project schedule.
– The cooperation felt easy. As a company we didn’t have to worry about whether development was progressing and we could trust that the goals would be achieved, he says.
The practical way of working was straightforward: a shared Slack channel was created, interim deadlines were set for different versions, and new approaches were tested. Instead of focusing primarily on technical implementation, ReactorEd was able to concentrate on user experience – how the tool should look and feel.
After a successful PoC, moving forward to full development was a smooth process. Today, the AI tutor is already in test phase use in four municipalities, reaching 1,000–2,000 learners.
– Our customers are teachers and their job is to teach. It makes little sense for them to spend time manually building repetitive practice exercises when those can be automated. That time is better used for actual teaching, Savolainen notes.
The Value of Expertise
In addition to developing the AI tutor, ReactorEd also received FAIR expert support for another tool. The company had developed an AI-based language proficiency assessment, which it was able to pilot in practice with the City of Espoo’s employment services.
The tool was designed for jobseekers whose native language is something other than Finnish. Running in a web browser, it assesses reading, writing, speaking, and listening skills. After roughly a 20-minute test, the jobseeker receives a PDF with results that can provide both the individual and employment officers with a clear understanding of language capabilities.
– Together with FAIR’s experts we were able to deepen and clarify the concept. Whether we had overlooked anything and what we could improve in the future, Savolainen says.
Even though technical implementation happened elsewhere, sparring with experts brought clear value. It was also beneficial for the company to have a credible expert organization like FAIR as a partner.
– I’ve spent ten years as a startup entrepreneur trying to break through walls. FAIR is connected in many directions, and their network can be highly valuable. Their expert brand can also add credibility to partner projects, Savolainen reflects.
Doing the Right Things at the Right Time
For many companies, the threshold to developing AI-based solutions can be high if the required expertise does not yet exist internally. And even if AI is already familiar, expert support can significantly accelerate a new initiative.
– Entrepreneurship and developing these kinds of solutions is often trial and error. If experts can help you avoid the errors and do more trials that really move things forward. It speeds up the whole process and saves resources, Savolainen states.
– Companies also need to be willing to accept feedback. You can’t treat your idea as perfect. An expert can help you see what might not work and how to improve things. But you have to be willing to listen, he continues.
According to Savolainen, the greatest benefit of working with FAIR was being able to make the right decisions from the very beginning.
– We gained early confidence that our idea could be turned into a working tool. We started with the right things and were able to execute efficiently. On our own, we might have done something roughly similar, but this collaboration helped us in the right direction from the start, he concludes.
