AI that works: Our take on what actually delivers

AI isn’t just the next tech buzz; it’s already changing how digital products are designed, built, and improved. But with all the hype, it’s easy to lose sight of where the real value is. What does useful AI actually look like in a business context? That’s the question we keep coming back to.

What we’re seeing

At Codemate, we’ve been building AI-powered solutions for years. Not for the sake of chasing trends, but because they help our customers solve real problems. We like to think of ourselves as AI pragmatists: more interested in what works than what wows.

At Nordic Business Forum 2025, our CEO Toni Piirainen and Director of AI, Data & Cloud Ville Lindfors shared what we’ve learned from working with companies like Netflix, Moi Mobile and beyond. Not just about tech – but about culture, decision-making and transformation.

Here’s a quick recap. To see the full presentation, scroll to the end for the recording!

We’ve helped companies:

• Make smarter decisions with their data
• Automate the right parts of their operations
• Create smoother, more human and personalised user experiences
• Build the right tools and frameworks so that not only developers, but also process owners can work with AI safely, transparently, and within auditable boundaries.
• Refine AI strategies so they are ambitious, responsible, and understandable across the organisation

We also use AI in our own work, from design to development, testing, and analysis. It’s not a magic wand, but it helps us move faster, work smarter, and spend more time on what matters.

“The best AI projects don’t start with tech. They start with a business challenge. That said, technology shifts create opportunities we shouldn’t ignore. When customers move early on those, they often help define the market,” says Ville Lindfors, Director of AI, Data & Cloud at Codemate.

“If we can’t tie it to a real outcome, we don’t build it,” he continues.

Our principle: AI helps people achieve more

AI should be useful, understandable, and responsible. That’s the standard we hold ourselves to, and it shapes every solution we build. In practice, responsibility means:

• Training models on your own data
• Business processes owners can define how things work, not just developers
• Monitoring performance in the real world, not just in testing
• Keeping humans in control, always

It’s easy to get carried away with what AI could do. But we’re more focused on what it should do. That means building tools that people can trust and ensuring AI remains a helper, not a black box.

Real stories from the field

So, what does this look like in real companies?

We’ve seen firsthand how even small teams can compete with bigger players by embedding AI into their core processes – and how success often depends more on mindset than on models. 

At Moi Mobiili, a lean Finnish telco, AI has become part of daily operations. What started as app development has evolved into a system where business people themselves can teach and manage AI tools. The result? Faster iteration, better transparency, and efficiency at scale.

Netflix is a different beast – but lessons are surprisingly similar. Our team helped them make Metaflow, their ML infrastructure platform, accessible beyond developers. Today, it’s open source, still evolving, and still part of what we do through our Outerbounds partnership.

These are not experiments. They’re working systems – and they’re changing how organisations operate.

In these cases, AI helped to:

• Fuel growth with new digital services
• Make customer experience faster, more personal, and more efficient
• Enabled teams to work smarter and scale operations

These stories show what’s possible when you combine strong technical foundations with a clear purpose. The best AI solutions are built on two pillars: they make life easier for people and deliver measurable value to the business. Our job is to make sure both happen – always at the same time.

What’s next?

We’re continuing to explore what AI can unlock – especially in collaboration with Google, Flutter teams and our other partners. We’re also running tailored workshops and building real systems that help companies move from AI talk to AI impact.

More about our work with Flutter

Watch the talk 👇

Originally published on the Reaktor Ecosystem blog. Updated for our own channels.