Sudipto Dasgupta is Global Head of Data, AI, and Automation Platforms at Aon.
The insurance industry is at a pivotal moment. Historically reliant on manual processes and legacy systems, insurers now face mounting pressures: rising operational costs, evolving risk landscapes and heightened customer expectations for personalized, digital-first experiences. That makes artificial intelligence (AI) a critical driver of enterprise transformation across underwriting, claims, brokerage, advisory and customer engagement teams.
According to McKinsey researchers, generative AI (GenAI) alone has the potential to unlock significant productivity gains globally, and insurers are among the sectors best positioned to benefit most.
Barriers To GenAI Adoption In Insurance
Data Siloes: Insurance organizations often operate with fragmented legacy systems, making it difficult to establish a single source of truth for AI models.
Process Variance: Insurance processes vary widely across product lines, regions and business taxonomies. This variation affects model grounding and can reduce the consistency of output from the AI models.
Regulation And Compliance: Compliance with GDPR, FCRA and emerging AI-specific regulations adds layers of complexity. In a highly regulated industry, transparency and explainability are essential for adoption.
Trusted Data: Key To AI-Driven Transformation
Given the prevalence of siloed data, establishing trust in enterprise data is essential. AI outputs can only be as accurate as their underlying data.
Centralized Master Data Governance: Master data across policy and transactional systems should be governed centrally. Consistency standards improve data quality and enable scalable model performance.
Global Taxonomy: Product, organizational and industry taxonomies must be defined and maintained centrally. A consistent taxonomy creates the foundation for integrating previously siloed datasets and enabling enterprise-wide AI use cases.
Data As A Product: Treating data as a product elevates data quality as a core design principle with clear ownership, life cycle management and user-centric design. This enhances trust in AI outputs over time.
Precision, Governance And Minimizing Hallucinations
Hallucinations occur when AI models generate plausible but incorrect information. In insurance, where errors carry financial and reputational consequences, reducing hallucinations is essential.
For business-critical processes like underwriting, claims processing and document review, the rigorous grounding of models and a precision-driven design approach can improve reliability. Ensuring humans remain at the center of these processes further strengthens trust, particularly for tasks requiring contextual judgment or regulatory oversight.
Defining AI principles around transparency, fairness, accountability, privacy and security should guide model design. Policies and controls should reflect regulatory requirements, with clarity around ownership and automated monitoring to ensure they scale effectively.
Measuring ROI And Prioritizing High-Value Use Cases
Use cases should be prioritized based on quantified outcomes. A design-thinking lens can help evaluate viability at the intersection of:
• Desirability: Does it solve a meaningful client or business need?
• Feasibility: Is the capability realistically achievable with current technology?
• Viability: Can it generate sustainable enterprise value?
For example:
• AI-driven financial reconciliation is highly desirable, but full automation is constrained by accuracy, auditability and financial-risk considerations.
• Ticket triaging is highly feasible given lower risk exposure, but may deliver limited ROI.
• Document intelligence in underwriting or brokerage offers both feasibility and strong commercial value. For example, automating the extraction of fields from submissions or carrier quotes can improve cycle time, enhance accuracy and accelerate time-to-value across placement workflows.
Strategic sequencing of these use cases helps maximize enterprise return.
High-Impact AI Use Cases Across The Insurance Value Chain
Claims Processing And Management: GenAI can process documents, freeform notes, text fields and images to automatically summarize, categorize and analyze claims information. This can accelerate claims processing, providing efficiency while improving the client experience. The global claims market is projected to grow significantly in the coming years.
Customer Service: Virtual assistants powered by large language models can support customers through both text and voice channels. This can help with filing claims, interpreting policies or accessing real-time information.
Document Analysis And Extraction: Insurance involves the exchange and processing of complex documents across customers, carriers and brokers. AI-driven document extraction can help process information from documents easily, reduce manual intervention and free up time for claims colleagues to deliver value-added services for clients.
Knowledge Management: Domain-specific chatbots can help provide specialized and trusted knowledge on topics like claims guidelines and policy management.
AI As A Catalyst For Reimagining The Insurance Industry
GenAI offers insurers an opportunity to leapfrog legacy constraints, accelerating processes, reducing friction and creating new value for clients. But lasting transformation requires three critical foundations: trusted, governed, high-quality data, robust AI governance and precision-first model design and clear, ROI-driven prioritization of use cases.
With these elements in place, insurers can unlock new frontiers of efficiency, redesign client experiences and operate with speed that was previously unattainable. Those who pair innovation with disciplined execution will be best positioned to scale AI responsibly, intelligently and with enduring strategic impact.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
