28.04.2026
Refox CEO Petter Lillhonga rides the crest of the AI wave and actively brings AI into the company's processes. The biggest benefit for the customer is the speed AI delivers.
— A customer can easily present us with an idea or a wish, and we can practically test
whether it works right away. What used to require several meetings and hours of planning can now
be demoed instantly with AI, Lillhonga says.
One of the most concrete benefits of AI is in how different needs can be served.
Companies use data in different ways. For one, a visual product view matters most; for another,
detailed analytics. Traditionally, this has meant compromises from a software point of view.
– When we built one tool for everyone, we had to consider every user. Now, instead, we can
test and build solutions tailored to a single customer without affecting anyone else.
In software development, testing has been one of the most labour-intensive phases.
– There's a model called test driven development, where you first write the tests
and only then the code. It's a very effective and almost perfect approach, but in practice it's
incredibly laborious, Petter Lillhonga says.
AI brings relief here too.
– AI can handle the tedious part. And when that process is used correctly, code quality
improves.
Errors don't disappear entirely, of course.
– We'll probably never reach a point where mistakes don't happen. But AI is correcting them
better and faster all the time.
Among other things, Lillhonga is experimenting in his own work with AI teams. That means he
builds entire "agent teams" where AIs specialised in different roles collaborate.
– A team might include, for example, a systems architect, a UI designer, and a translator.
One acts as team leader and the others handle their own assignments.
Petter describes the experience as slightly bewildering.
– They talk to each other, write instructions and documentation for one another. The first
times you see it, it actually feels a bit unsettling when you realise what AI is capable of.
Human work is still needed.
– My role is essentially to supervise. I get a report on what the agents have done and I
steer them towards the next stage.
The fact that Refox actively uses AI shows up very concretely for the customer. Troubleshooting
is faster, and at best, the system fixes the situation automatically.
– If AI notices a bug, it can even fix it so the customer can continue working straight
away.
One of the biggest changes concerns who can leverage data. Previously, putting data to use
required technical expertise. That is now changing.
– Customers have the opportunity to build tools for themselves that let them read and
interpret data without years of IT training.
Petter divides AI users roughly into four levels based on how deeply they use it.
At the lowest level are those who have perhaps heard of AI or tried it occasionally, but don't
use it in their daily lives.
The next level includes users who leverage AI for individual tasks, such as producing text or
solving problems.
On the third level are companies that have already integrated AI into their own processes and
continually consider it when designing new ways of working.
At the top level are the pioneers, who don't just use AI but actively direct it, build new
solutions on top of it, and clear the path for others by showing what is possible.
Refox is on the third level — and partly even on the top level.
Petter wants to encourage companies to think about things in a new way.
— Previously, many ideas went unrealised because they were too big, too expensive, or too
complicated. But what once felt impossible is very easily accessible today, Petter says.
– I would like people to dare to think more often that the impossible is possible, Petter
Lillhonga encourages.
Tell us about your business. We'll do an assessment and show you how Refox solves your specific challenges – directly from the developer, no middlemen.
