‍

From concept to code: AI in practice

Published

August 14, 2025

Author

Mads Hofman

CEO & Advisor

You’re probably used to AI talk drifting into theory and predictions, but our approach is different. We use this technology in our day-to-day work and, in selected projects, we’ve seen a significant boost in efficiency as a result.

We also like to share our findings on how AI can ease your daily workload through webinars and our newsletter. On this page, you’ll find examples of where we see AI evolving right now, and how we’ve put it to work in our own development.

Your virtual assistant

The custom chatbot we built for Dao instantly streamlined their operations, freeing up 20% of their customer support capacity.

Staying sharp

We work continuously with the technology to understand what actually works in practice. If you’re looking for practical, distilled insights into AI developments, stay sharp through our newsletter. Here are two relevant examples of the kind of updates you can expect.

‍AI-first tools

We’re seeing tools that are built around AI from the start, not just supported by it. The question is no longer how AI can assist, but how much of the process it can handle. We use platforms like Bolt.new, Cursor, Lovable and Replit as starting points for our explorations. They provide a strong foundation for building an AI agent that can collaborate closely with a real developer throughout the process.

Conversational Shopping

In April 2025, ChatGPT launched a shopping assistant that understands natural language. You can ask for something like "a quiet vacuum for a small apartment with wooden floors" and get tailored suggestions without logging in. It's a clear shift from search to conversation.

How we're using it

These projects show what AI can do in practice and they’re helping us learn where the technology fits best.

Custom chatbot
With dao, we developed a custom chatbot to ease the load on their customer support team. It now handles a large share of incoming questions around the clock, seamlessly integrates into the customer experience, and frees up the team to focus on everything else, while customers get faster answers.
Activity planning enhanced by AI
In our collaboration with Det Danske Spejderkorps, we used AI to support activity planning for scout leaders. The goal was to make it easier to build coherent activity plans across a wide range of existing content - while staying true to the core of the scouting method, where children learn leadership through experience. AI helped bring structure, overview, and relevance to what was previously a dense and scattered pool of materials. 
Offer generator
‍For Byggetilbud, we built a system that generates renovation offers for private customers. It was developed with an AI-first approach to explore how far we could go beyond using AI as just a support tool. We used the platform Replit to prototype with AI from the beginning. It let us test ideas quickly and explore AI as an active development partner. Which resulted in an efficient offer process, that quickly tailors to the customers individual needs.