Building features in hours, not days

Nikolas Kohvakka, principal consultant at Kodan, showed in his presentation how AI tools accelerate development, demoing a real product development process where a geolocation-based feature was added to the product with AI. Manual coding of this feature would have taken two to three days, but today, using AI agents and tools like Cursor, we can design, build, test, and document that same full-stack feature in under an hour.

During his presentation, Nikolas walked us through the whole AI-powered development process and gave the coding agent the task of creating the geolocation feature. At the end of our event, Nikolas showed us the fully working feature in all its glory.

A solid tech foundation is non-negotiable

Tommi Jalkanen, the co-founder and senior consultant of Bruvo, shifted the focus from AI models to the data architecture behind them, asking: if we had a perfect AI model tomorrow, what would stop us from deploying it next month? The bottleneck is rarely the model; it's scalability, compliance, and data hosting.

To build a reliable and valuable foundation for AI solutions, he urged the audience to take a problem-first approach: "Using AI" is not a business case. Solve actual business problems — don't just chase technology.

AI Governance: Managing your "agent armies" safely

Miika Kankare, CEO of Kanto Company, ended the afternoon with an experience-backed presentation about AI governance. With organizations deploying "agent armies," governance is essentially modern risk management. Unlike rigid traditional automation, LLM-powered agents are adaptive, which means they can make unexpected decisions or hallucinate.

Miika highlighted two rules for safe governance:

• Never run agents under a developer's personal credentials. Give each agent its own identity and minimum access rights.

• Use a central platform to maintain auditability and manage costs, especially with the EU AI Act coming.

When auditors knock on your door, "the AI did it!" won't save you. A human must always remain accountable.

The future is already here

AI is no longer futuristic or speculative. It's a practical tool for solving real business challenges, even if it's still rapidly evolving. And success with AI requires clean data, secure architecture, safe governance, and a senior team to connect the dots, not just randomly bought AI licenses and scattered test projects.

If you want to safely accelerate AI-powered development, clean up your data foundations, or build reliable AI workflows, let's talk!