Artificial intelligence (AI) is moving rapidly from experimentation to execution in the property/casualty insurance industry. From AI-driven claims estimation to predictive underwriting and fraud detection, insurers are realizing tangible value in faster decisions, improved accuracy and stronger customer experiences. Yet, for every success story, there are stalled initiatives and missed expectations.

The difference often comes down to one critical factor: data readiness.

Most P/C insurers operate within a patchwork of legacy core systems (policy, billing and claims), bolstered by spreadsheets and specialized departmental applications. These siloed environments create inconsistent definitions, incomplete records and limited visibility across the enterprise. When AI models are trained on fragmented or poor-quality data, they inevitably underperform—eroding trust and slowing adoption.

Data First, AI Second

Rather than rushing to build narrowly focused AI models, insurers and their partners must first strengthen their data foundation. Without ensuring data quality, lineage and accessibility, no AI initiative will achieve success across the enterprise.

Insurers increasingly recognize the need for a modern Data Lakehouse—a unified environment that consolidates structured and unstructured data from core systems, underwriting and external sources. This conformed, auditable platform, aligned with insurance data standards such as ACORD, enables the creation of a single source of truth that fuels all downstream analytics and AI applications.

Related article: Demystifying the Data Landscape: Lake, Warehouse and Lakehouse Explained

The benefit is clear: Organizations can minimize the risk of failed AI projects, accelerate time to value and maximize return on data investments while maintaining transparency and governance.

The Medallion Architecture: From Raw to Refined Intelligence

At the heart of the Data Lakehouse concept lies the Medallion Architecture, which organizes data into layered stages—Bronze, Silver and Gold.

• Bronze captures raw, unprocessed data directly from operational and external systems, preserving lineage and ensuring full traceability.
• Silver transforms, cleanses and standardizes this data, applying consistent business rules to create a reliable, insurer-wide view of customers, exposures and transactions.
• Gold provides curated, analytics-ready datasets that support underwriting, pricing, reserving, fraud detection and loss analytics.

This tiered approach moves insurers away from ad-hoc data pipelines toward a disciplined data-as-infrastructure model—a living, trusted foundation where AI models can confidently consume accurate, governed data.

Azure Fabric: The Next Step in Unified Data and AI

Microsoft’s Fabric platform brings this vision of data infrastructure to life by unifying data engineering, analytics, business intelligence (BI) and AI within one SaaS environment. Fabric’s OneLake concept enables all enterprise data—structured or unstructured—to coexist in a single logical lake, accessible through consistent governance, security and policy controls.

For insurers, this means claims adjusters, actuaries and underwriters can access harmonized data through familiar tools such as Power BI or directly query it using natural language through Azure OpenAI Service. AI becomes part of daily operations—not a separate experimental initiative.

By integrating Lakehouse principles and AI-native capabilities, Microsoft Fabric turns data infrastructure into a living intelligence layer, where ingestion, transformation and modeling happen seamlessly. It bridges the modernization gap for insurers, allowing them to advance analytics without replacing every legacy system.

Data Mesh: Scaling Data Ownership Across the Enterprise

While the Lakehouse and Fabric emphasize technological unification, the Data Mesh model extends the concept into organizational design. In a Data Mesh, data is treated as a product, owned by the business domains that generate and understand it—claims, underwriting, finance and operations—while adhering to shared governance and interoperability standards.

This federated approach allows insurers to decentralize innovation while maintaining global consistency. Local domain teams can quickly build and deploy models that address their unique needs, while the enterprise ensures data quality, lineage and compliance across regions and lines of business.

When combined with a Lakehouse foundation, Data Mesh architecture delivers both autonomy and alignment, scaling AI safely and efficiently.

Enabling Generative AI

Early AI efforts in insurance focused primarily on automation and pattern recognition. Today, the industry’s attention is shifting to generative AI (GenAI), which allows users to explore and interact with data through natural language prompts.

Imagine an underwriter asking: “What are the top five drivers of loss ratio deterioration in commercial auto this quarter?” Instead of waiting for a specialized analytics team to generate a report, that underwriter receives an immediate, contextual and data-backed explanation. Actuaries, claims leaders and executives gain similar access to insights—without requiring deep analytics expertise.

However, this conversational intelligence only delivers value when GenAI tools are grounded in trusted, well-governed data infrastructure built on Lakehouse, Fabric and Mesh principles.

Complementing, Not Competing

The ecosystem’s greatest strength lies in cooperation. Data and analytics platform providers must work alongside AI specialist firms such as Tractable (image-based claims), Shift Technology (fraud detection) and Cape Analytics (geospatial risk). These AI models thrive only when the underlying data is clean, complete and explainable.

No algorithm can outperform the quality of the data it consumes.

A Strategic Edge

As insurers transition from proofs of concept to enterprise-wide AI adoption, trusted data will define the next generation of industry leaders. By embracing Data Lakehouse Medallion Architecture, leveraging Microsoft Fabric’s unified analytics and AI capabilities, and adopting Data Mesh governance, insurers transform their data into an enterprise-wide infrastructure asset.

Those who treat data as infrastructure—not an afterthought—will be best positioned to harness AI confidently, responsibly and at scale. The future of AI in insurance belongs to those who build trust in data first and innovation second.