Fremtind — AI-driven predictive car insurance
Product and Design Lead, full responsibility for service design, interaction design and UX. Scrum Master and interim PO for a development team of 9. End-to-end from concept through iterative implementation.
Fremtind (formerly Sparebank1 Insurance) had a car insurance portfolio losing competitiveness against new fully digital insurance solutions. The challenge: design tomorrow's car insurance product from scratch — built on big data and machine learning.
The concept was fundamentally AI-driven: instead of historical risk models, the solution used machine learning to identify individual driving patterns and behaviours in real time, then automatically build a personalised insurance price based on how that specific customer actually drives. Better driving behaviour meant a lower premium — and the app was the interface that made this visible and actionable to the user.
The core UX challenge was designing for AI-generated, personalised output at scale. Every user saw a different insurance price, different behavioural insights, and different recommendations — all generated by the ML model. My job was to make that AI output feel trustworthy, fair, and motivating rather than opaque or arbitrary.
A key design constraint: all meaningful user interaction had to happen outside of driving time. I mapped touchpoints carefully to ensure the app engaged users before and after driving — never during.
The solution also included gamification mechanics to encourage safer driving behaviour, reducing mobile phone use while driving by 55%.
Results