As insurers look to be more agile, data mesh strategies take centerstage

In an era of business where every industry requires its stakeholders to be capable of quick pivots and sharp turns, siloed information that ultimately slows decision-making can be the ultimate vulnerability. In this way, data may just be the ultimate disruptor – a fact that the insurance industry knows all too well.

As data volumes continue to increase alongside a correlating number of business requests, modern insurance data leaders face a nuanced set of challenges. Ever-shifting domain-level business logic and architectures add to the workload of overwhelmed central data teams, which results in difficult, misaligned metric reporting and declining data reliability. Accelerated demand in AI-enabled innovations has recently compounded these issues, prioritizing the need for new capabilities that require even more robust data foundations. 

Enter data mesh. With its emphasis on decentralized, domain-oriented data ownership and architecture, data mesh provides a potential answer for overmatched, out-manned businesses. It can empower individual business units to manage their own data domains and address many of the common challenges facing carrier data environments, all allowing for faster decision-making, and more agile environments that evolve with the data rather than fighting to continuously force ubiquitous harmonization. 

The data mesh debate

This is not to say that there is a consensus that data mesh is a universal solution. Stakeholders are currently waging an open debate across the industry of centralization versus federated data strategies.  Proponents of centralization continue to assert its effectiveness in driving operational efficiencies, enhancing analytics effectiveness, and enabling governance crucial to data security, privacy, and regulatory compliance.  Despite these benefits, the core problems that data centralization so often fails to address are the pragmatic realities of many enterprise data ecosystems.

For starters, there are many instances where most of the business day-to-day use and management of a firm’s dataset are based on the “fit-for-purpose” needs of its constituent area.  Harmonizing these data sets with centralized enterprise data faces increasing challenges as shifts to data definitions, schema, and architecture require constant central data team efforts.

The human political element plays a significant role here as well, as local data owners push back on relinquishing control over domain-specific data assets to centralized data governance authorities. These domain data leaders often cite the diminishing returns and significant effort of central data team engagement. Additionally, data silos and fragmentation often occur inorganically as in the case of merger or acquisition scenarios. Isolated data sources requiring long data transformation and integration timelines can become stumbling blocks to the realization of combined organizational benefits.

The unified need

Despite these criticisms, the need for unified, integrated data solutions at the enterprise level remains.  To deliver a cohesive data strategy, seamless data access and sharing requires the breaking down of data silos. This is where data fabric tools with their focus on orchestration, contextual layering, and metadata management are important elements to add to the equation.

Data fabric introduces an intelligent semantic layer that orchestrates disparate data sources, applications, and services into a unified and easily accessible framework. Enabled via a data integration hub, the data fabric architecture connects, organizes, and manages data, providing a consistent view across the data ecosystem. End-to-end data fabric capabilities that handle master data management (MDM), data matching, real-time data integration, data quality, and observability can all be implemented without wholesale replacement of current tech stacks.  Data fabric applications leverage leading AI-enabled capabilities to prepare and deliver data for deeper insights and efficient capability development at the enterprise level.

Interestingly, data leaders should not see the federated data mesh strategy set versus more centralized data fabric architectures as an “either-or” dilemma. On the contrary, data mesh methods are focused at their core on data ownership, leveraging micro-services-inspired governance methods, and emphasizing local domain ownership and accountability. This makes data mesh work in cases where a significant percentage of data use cases are local and data structures are relatively dynamic in nature.  In those instances where enterprise data integration is required to enable cross-domain capabilities, data fabric, and digital integration hub architectures enable flexible interoperability, rapid data processing, scalability advantages, and cohesive data governance across a complex enterprise data environment. For insurance, carriers that seek to enable the benefits of federated data mesh strategies while also addressing the need for seamless integration of disparate data sources – incorporating both a data mesh strategic approach with data fabric architectural toolset may allow carriers to enjoy benefits from both sides of the strategic spectrum.  

It should be noted that implementing a blended data management strategy is not without potential pitfalls. Implementation complexity is a leading strategic risk as design, solution implementation, and maintenance all introduce challenges, particularly as it relates to existing legacy systems and architectural restructuring. Integrating new technologies into an existing data ecosystem is equally complex, with compatibility issues and the need for substantial data migration introducing increasing potential for operational disruptions.  

A hybrid future

No data transformation journey is without its own unique set of perils and pitfalls, but solving this complex problem of mounting data volumes requires innovative thinking. The advancement of AI-enabled data fabric tools and architectures allows data leaders to reassess long-held philosophies and take a pragmatic approach.

A combination of data mesh and data fabric allows Insurance organizations to incorporate resilient, flexible governance structures that are sustainable by the organization and that deliver on the needs for innovation and growth. For insurers looking to strike the right balance, embracing a hybrid strategy approach serves as an attractive path for the industry’s data-led future.

To learn more, visit us here.

About the author:
Karl Canty is vice president, life insurance, annuities, and group benefits analytics at EXL, a leading data analytics and digital operations and solutions company.