Data Mesh Unlocks $40M in Revenue, Empowers Teams to Innovate

Together we are:

Building a scalable Data Mesh architecture
Enabling secure, federated data-sharing
Reducing churn through predictive retention modeling
Applying AI to improve underwriting and retention
Creating a blueprint for global data transformation
Unlocking a projected $40 million annual uplift

The Client

This global insurance group has operations in more than 40 countries. For this project, we worked with the group’s shared services organization.

The Challenge

The client was facing a complex challenge: it needed to enable each business to manage and analyze its own data while maintaining global consistency, trust, and compliance.

Each company in the group operates differently, with unique customer bases, regulatory obligations, and data models. Those differences led to duplicated systems, inconsistent reporting, and limited visibility. Sharing data across entities was slow and risky, especially when it involved sensitive customer information.

They needed a solution that could balance local autonomy with global oversight: a modern architecture that encouraged innovation but maintained governance, privacy, and control.

The Journey

To show the value of a new approach, we co-designed two pilot use cases where better data could make an immediate difference.

Underwriting Efficiency

Underwriters manage thousands of quotes each day but can only process a fraction of them. A new propensity-to-bind model helps identify the quotes most likely to convert, enabling underwriters to focus on the 20% of submissions that generate most of the revenue. The result: faster decision-making and a clear uplift in underwriting performance.

Policy Retention

While the client already retains 91% of policyholders, the 9% churn represents nearly $360 million in lost premiums each year. By applying machine learning to customer behavior, broker interactions, and product data, they can now predict renewals and proactively address at-risk accounts. Even a 1% improvement in retention equates to a projected $40 million annual uplift.

Both use cases required the secure exchange of data across multiple business lines—with strong privacy, compliance, and data-residency controls in place.

To enable a balance of autonomy and oversight, we implemented a Data Mesh architecture built on Databricks and Unity Catalog. The model gives each business unit freedom to innovate while ensuring consistent governance, lineage tracking, and compliance.

Central data registry: A Databricks workspace acts as the integration hub and authoritative registry, with Unity Catalog enforcing policies and permissions.
Domain workspaces: Each business manages its own Databricks environment for pipelines, analytics, and AI models — governed by shared global standards.

This design supports granular access control, regional data-residency requirements, and cross-domain collaboration, creating a single, trusted framework for secure data-sharing across the group. Also, aggregated or anonymized data can be shared across businesses where full visibility isn’t permitted, ensuring compliance whilst still encouraging collaboration. Central dashboards now monitor data quality, freshness, and schema consistency across workspaces, strengthening trust in shared data.

The co-creation model enabled client teams to build new internal capability and sustain the Data Mesh independently after launch. The partnership also ensured rapid alignment, sustainable delivery, and a scalable foundation for future innovation.

The Outcomes

This initiative is reshaping how data is governed and used across the enterprise. It enables the organization to:

Unlock $40 million in potential annual revenue uplift through retention and underwriting models.
Empower business units with autonomy while maintaining governance.
Establish federated oversight with consistent compliance controls
Enable secure, trusted data sharing across business lines and geographies.
Strengthen collaboration and embed a culture of data excellence across the group.

With the scalable model now proven, the client is preparing for global rollout: building a connected, compliant, and insight-driven data ecosystem that combines local agility with enterprise control.