Mitsui Sumitomo Insurance

Mitsui Sumitomo Insurance, the core non-life insurance arm of Japan's MS&AD Insurance Group with ¥1.6 trillion in net premium income, faced a structural data problem common in large P&C carriers: more than 180 terabytes of operational data accumulated across disconnected data warehouses and marts. Because agents are the primary distribution channel in Japan's insurance market, their ability to recommend the right product at the right moment — particularly during policy renewals — directly drives premium income. Fragmented data prevented the company from building the unified analytics layer required to support over 34,000 domestic agents with AI-powered guidance, leaving significant renewal upsell potential unrealized.

To address this, Mitsui Sumitomo Insurance deployed Teradata VantageCloud on AWS as a consolidated enterprise data platform, decommissioning legacy siloed warehouses and centralizing internal data alongside group company, agent, and business partner data. On this foundation, the company built MS1 Brain — a sales support platform purpose-built for its domestic agent network. The platform applies predictive machine learning to forecast customer demand and product fit, then surfaces Next Best Action (NBA) recommendations that guide agents toward the most suitable supplemental coverage for each policyholder at renewal. Additional capabilities include personalized customer-facing videos explaining coverage options and business management dashboards for agents. The massively parallel processing architecture of VantageCloud enabled high-performance query execution across complex workloads while maintaining SLAs and delivering predictable costs.

The deployment of MS1 Brain's NBA functionality produced a measurable step-change in renewal performance. Key outcomes include:

• ~250% increase in the rate at which policyholders add supplemental coverage during policy renewals
• 34,000+ domestic agents actively using the platform across Japan
• Scalable query performance maintained across large, complex workloads without unexpected compute cost spikes

Beyond the headline metric, the initiative shifted analytics from a specialist function to a company-wide capability. Mitsui Sumitomo is now extending data access to all employees — not just analysts — through digital training programs designed to democratize data-driven decision-making across the organization.

• Data unification precedes AI value: Consolidated infrastructure is a prerequisite — NBA recommendations are only as good as the breadth and quality of data feeding them.
• Renewal moments are high-leverage AI touchpoints: In P&C, the renewal interaction is where AI-guided upsell has the greatest impact on premium income.
• Agent augmentation outperforms agent replacement: Embedding AI guidance into agent workflows, rather than bypassing agents, compounds the value of a large distribution network.
• Workload SLAs are a vendor selection criterion: Cloud platforms with limited workload management introduce cost unpredictability at enterprise scale — evaluate this before committing.
• Design the platform for future users, not just current ones: Expanding from specialist analysts to all 34,000+ agents requires scalability built in from the start.