Solera's Qapter auto-authorizes 50% of damage claims with computer vision, cutting estimation costs by half

Insurance claims departments were slow and frustrating for customers, relying on physical inspections by claims adjusters. Prior attempts by insurers to automate collision damage assessment with computer vision failed to scale. Solera needed to modernize its Qapter claims workflow platform to meet modern customer expectations for speed.

Solera rebuilt Qapter using Google Cloud's Vision API and TensorFlow to process touchless claims from a single photo. The system uses OCR to capture license plates and VINs, custom ML models for vehicle make/model recognition and damage identification, and Cloud GPUs/TPUs to accelerate model training. Qapter compares damage images against a large proprietary repository to estimate repair scope, parts needed, and final cost.

Within months of launching in France and the Netherlands in 2020, Qapter could auto-authorize 50% of damage claims, reducing estimation costs by nearly half. The touchless process eliminated the need for in-person adjuster visits, benefiting drivers, insurers, and repair shops — and proved especially valuable during COVID-19 by minimizing human contact.

• Narrow AI scope to a specific workflow step (damage identification) rather than the entire process to avoid scaling failures.
• Augmenting an existing proprietary data lake (automotive images and parts catalogs) with cloud ML is more effective than building from scratch.
• A touchless claims workflow creates compounding benefits across the entire value chain — cost, speed, and safety.