How Insurer Aviva Migrated 1.3PB of Siloed Data to Become "AI-Ready" In 7 Months

Why It Matters

A centralised, AI‑ready data foundation gives Aviva faster time‑to‑market for digital products and positions it ahead of competitors still grappling with fragmented data.

Key Takeaways

• Aviva shifted 1.3 PB from Oracle to Snowflake in 7 months
• Centralised architecture now powers AI-driven underwriting and claims
• Enables insurance purchases directly through ChatGPT interfaces
• Migration reduced data silos, accelerating product development cycles

Pulse Analysis

Moving 1.3 petabytes of data in under eight months is a logistical feat for any insurer, especially when the source environment is a legacy Oracle Cloud setup. Aviva chose a lift‑and‑shift approach, replicating raw tables into Snowflake’s cloud‑native warehouse, which eliminated the need for extensive re‑engineering. Snowflake’s elastic storage and compute separation allowed the team to scale resources on demand, keeping migration windows short and cost predictable. The rapid timeline also reflected disciplined project governance, automated testing, and a clear data‑ownership model.

With the data now consolidated, Aviva gained a single source of truth for underwriting, claims, and customer analytics, a prerequisite for reliable AI models. The unified platform feeds natural‑language processing engines that power chatbots capable of handling policy inquiries, while a separate model integrates with OpenAI’s ChatGPT to enable end‑to‑end insurance purchases through conversational interfaces. By decoupling data storage from processing, the insurer can experiment with new algorithms without disrupting production workloads, shortening the innovation cycle from months to weeks.

Aviva’s migration signals a broader shift in the insurance sector, where legacy data silos have long hampered digital transformation. Competitors are now evaluating similar cloud‑first strategies to unlock AI potential and meet rising consumer expectations for instant, personalised service. The cost savings from reduced infrastructure sprawl, combined with revenue upside from AI‑enabled products, are expected to improve margins in a low‑interest‑rate environment. As regulators become more comfortable with AI‑driven underwriting, insurers that have already built an AI‑ready data foundation will likely capture the early‑mover advantage.