Leading insurer unlocks CHF 25M with a scalable AI pricing engine

The challenge: Fragmented pricing operations hindered scale, speed, and ownership

Our client struggled with low operationalisation of existing pricing streams, which not only led to long time-to-market but also made it difficult to scale similar pricing initiatives. Past collaborations lacked sufficient knowledge sharing, resulting in poor maintainability and ongoing reliance on external teams. Disconnected data, IT, and business teams further limited knowledge sharing and alignment.

How we helped: AI pricing-engine built on Vertex AI

Visium partnered up with the client, to design and deploy a robust AI pricing-engine workflow using Vertex AI. This enabled scalable and reproducible ML processes that integrate seamlessly with the client’s existing ecosystem. The solution includes:

• A robust, end-to-end ML workflow that integrates with the client’s wider application ecosystem

• Modular architecture allowing multiple models with distinct logic to be served via a single endpoint

• Embedded monitoring, alerting, and autoscaling, ensuring stability and reducing post-deployment manual intervention

• A unified model registry and versioning strategy supporting transparent lifecycle management

We established strong development and deployment environments, laying the foundation for future use cases. Close collaboration across product, data, and IT teams, along with ongoing knowledge transfer, helped break down silos and ensure lasting alignment.

Solution: Scalable, stable & strategically transformative

Designed for scalability, our solution supports rolling out new models to a shared endpoint with minimal maintenance overhead. It is designed not just to solve today’s problems but to enable pricing innovation at scale.

• CHF >25M/year bottom-line impact on-track through continuous monitoring, with no increase in customer churn.
• 6 models served through a single endpoint with minimal maintenance overhead
• <4 weeks average deployment time for new models, dramatically improving time-to-market
• The pricing team is now positioned as champion for MLOps, encouraging other teams to adopt the same best practices.

This AI pricing-engine is now a strategic asset, delivering measurable business value while enabling ongoing model innovation with minimal friction.

‍