Opterrix Cuts Hail Analysis From 10 Hours to 10 Seconds

Severe weather is becoming both more frequent and more intense, placing enormous pressure on insurers to assess and respond quickly. Hailstorms pose a unique challenge – they are localized, destructive, and unpredictable. Engineers were consuming up to 10 engineering hours per case manually digging through data to find matches – a slow, error-prone process that couldn't scale. Opterrix's lean team lacked the internal bandwidth, production architecture expertise, and implementation playbook to move quickly with GenAI.

Opterrix engaged DoiT as a strategic partner to bring generative AI into production. DoiT helped develop a solution that utilized GenAI to analyze images of hailstones and storm metadata, identifying similar historical events based on geospatial proximity and storm patterns. Leveraging Vertex AI and multimodal embeddings, the system transforms storm imagery and sensor data into searchable, vectorized intelligence in seconds. DoiT also optimized Opterrix's Google Cloud environment for performance and cost, implementing usage-based guardrails and introducing Cloud Intelligence™.

From the outset, DoiT approached the engagement not as a vendor, but as a strategic extension of Opterrix's team. The accelerator program – designed to bring projects to life in a structured but flexible way – provided the perfect framework. Early discovery sessions focused on understanding Opterrix's architecture, constraints, and goals in detail. DoiT didn't try to apply a one-size-fits-all model, they customized everything to Opterrix's business and platform – from the first workshop to the final deliverable.

Previously, Opterrix engineers had to manually compare data and imagery from past hail events – a slow, inconsistent process and highly dependent on individual expertise. DoiT helped Opterrix develop a solution that utilized GenAI to analyze images of hailstones and storm metadata, identifying similar historical events based on geospatial proximity and storm patterns. Leveraging Vertex AI and multimodal embeddings, the system understands visual and contextual features, transforming storm imagery and sensor data into searchable, vectorized intelligence in seconds.

The AI-powered hailstorm matching module replaced a labor-intensive, 10-hour process with an automated solution that delivers results in under 10 seconds, freeing significant engineering hours and accelerating insurer response. This leap in efficiency now allows Opterrix to respond faster to weather events, giving insurers near real-time intelligence to expedite claims processing, mobilize resources and reduce losses. By offloading the architectural design, model implementation, and infrastructure optimization to DoiT, Opterrix compressed R&D cycles by 50% and increased delivery confidence.

Beyond technical deliverables, the project elevated Opterrix's internal capabilities. Engineers gained hands-on experience with generative AI, understood how to integrate it into production systems and learned how to maintain governance in a regulated environment. The AI module is now a cornerstone capability for Opterrix and a launchpad for further innovation. But even more importantly, the team is now equipped to extend that innovation independently. This wasn't just about shipping a model, it was about transferring the skills, mindset and infrastructure needed to embed AI across their platform.