PZU SA

PZU SA, Poland's largest insurer and the leading financial group in Central and Eastern Europe, handles approximately 500,000 automotive claims annually — a volume that places enormous pressure on consistent, timely damage assessment. With 90% of cases requiring on-site physical inspection, PZU's network of 400 field assessors was the backbone of the claims operation. The inherent challenge with large distributed field teams is estimating variability: individual assessors apply judgment differently, producing inconsistent damage estimates, longer cycle times, and a higher rate of supplementary quotes that add cost and delay claim resolution across the country.

PZU partnered with Solera to implement Qapter Intelligent Estimating, Solera's next-generation AI estimating platform powered by computer vision. Rather than replacing the existing claims infrastructure, Qapter was integrated directly into PZU's established workflow as an upgrade path from the Qapter platform already in use — minimizing disruption and reducing the deployment burden. The system enforces PZU's own business rules within the estimating process, ensuring that AI-generated assessments align with company standards. A structured pilot validated the solution against three defined success criteria before full rollout. Within weeks of deployment, all 400 mobile field assessors were using the AI estimating capability on-site, with no significant workflow overhaul required.

The pilot exceeded PZU's expectations across all three evaluation pillars: accuracy, ease of use, and rollout effort. Key outcomes include:

• 400 field assessors equipped with AI estimating across Poland
• 500,000 automotive claims per year now processed through a standardized AI-assisted workflow
• 90% on-site assessment rate maintained, now with consistent, rules-based damage estimating rather than manual judgment alone
• Reduction in cycle times and supplementary quotes as estimating quality stabilizes across the field

The implementation marked PZU as the first Polish insurer to standardize AI damage estimating at national scale, setting a new operational benchmark in the Polish P&C market.

• Build on existing infrastructure: Integrating AI as an upgrade to a familiar platform, rather than a full replacement, was the primary driver of rapid adoption across 400 assessors.
• Define success criteria before the pilot: PZU's three-pillar framework (accuracy, ease of use, rollout effort) enabled objective evaluation and clear stakeholder alignment.
• Embed business rules in the AI layer: Configuring the system to enforce insurer-specific rules ensures AI output stays within acceptable estimating parameters, not just generic model predictions.
• AI scales consistency, not just speed: The primary value for large distributed field teams is reducing variance in estimate quality, which directly reduces downstream rework and supplementary quotes.