Data accuracy is critical
Reliability and forward thinking are core values of global insurer Zurich, with a long legacy of trust. Kubrick provided a custom solution for its complex reporting requirements to achieve the reliability and efficiencies required as well as setting them up for future development.
3 weeks
<1 min
100%
- 01
The challenge
Zurich aimed to address report ingestion challenges and had evaluated the best options for a scalable solution to meet their complex bordereaux requirements. It concluded that a custom solution would be required to address complexities, ensure regulatory compliance and provide flexibility, scalability and observability across operations.
The solution would need to:
- Close gaps in financial visibility and management.
- Connect data sources, formats and tools to improve efficiency and consistency.
- Remove manual interventions to reduce costs, delays and financial discrepancies.
- Increase confidence in data integrity and compliance monitoring.
- 02
The approach
Kubrick designed and implemented a solution with flexibility, scalability and advanced observability.
- Provided a Streamlit user-friendly interface: Allows business users to configure validation and transformation rules, upload files, and easily manage ingestion logic. This minimized reliance on IT teams for day-to-day operations.
- Built metadata-driven Snowflake pipelines: Using Snowpark, we employed a fully metadata-driven pipeline to process, transform, and load data according to predefined rules. This architecture allowed for seamless adjustments to ingestion logic as data standards evolved.
- Integrated monitoring and auditing: We implemented comprehensive logging and traceability using Snowflake Event Tables, enabling end-to-end monitoring of ingestion progress, error handling, and data transformation activities. An intuitive Streamlit UI was provided to track file progress and review logs for quick issue resolution.
- Minimized error handling and resilience: Erroneous records were captured in a rejected records table for easy review and reprocessing. The solution’s dynamic schema handling allowed it to adapt to new or altered data structures without requiring extensive rework.
- Provided Streamlit-based file upload UI: An additional tactical interface for manual file uploads, ensuring flexibility and continuity even during process exceptions.
- Ensured robust testing and CI/CD: Delivered with unit tests and a CI/CD pipeline, enabling continuous integration, automated testing, and rapid deployment.
- 03
The impact
- The rewards scheme reached 9m customers with a potential of £254m in incremental sales.
- Customer attributes used to personalize marketing campaigns increased from 40 to 125.
- The improvement in data quality and sources enabled the creation of C-suite level dashboards, improving strategic and tactical decision making.


