What does it take to introduce AI into a system handling some of society’s most sensitive data?
In this episode Sebastian speaks with Magnus Skebäck, CEO of Pulsen Omsorg AB, about the company’s journey from a legacy monolith to microservices and, more recently, towards AI-native products in Swedish social welfare.
Magnus shares why the earlier architectural transformation became an unexpected foundation for AI, and why adopting AI in a highly regulated environment turned out to be far more than a technology project. Legal requirements, privacy, security, documentation and customer trust shaped everything from the choice of infrastructure to the way data could be processed and stored.
The conversation also looks at how Pulsen Omsorg chose real operational problems instead of experimental AI pilots. One early use case focused on reducing thousands of hours spent documenting meetings and care activities, with success measured through time saved and improved quality. Magnus also explains why AI is changing the role of developers, why some of his team deliberately work without AI from time to time, and why moving too quickly can damage trust in ways that are difficult to reverse.
Key topics:
- Why modern architecture became a foundation for AI adoption
- How to identify AI use cases based on real customer problems
- Why compliance and privacy shaped the technology itself
- How AI is changing the skills developers need
- Why trust and gradual adoption matter in sensitive environments
⏱ Timestamps:
00:00 – Introduction
06:15 – From monolith to microservices and preparing for AI
18:54 – Measuring real value from AI adoption
27:35 – Why 70% of the challenge was legal
38:27 – Trust, responsible adoption and what comes next
🎙 Guest:
Magnus Skebäck, CEO of Pulsen Omsorg AB
LinkedIn: https://www.linkedin.com/in/magnusskeback/
🎙Host:
Sebastian Dzieniak
LinkedIn: https://www.linkedin.com/in/sebastian-dzieniak/
