QBE Insurance

QBE Insurance, a global insurer with roots dating to 1886, processes enormous volumes of claim forms and invoices across its worldwide operations. Claims volume spikes unpredictably following weather events, disasters, and other loss events, placing acute pressure on back-office document workflows. Manual processing was slow and labor-intensive, creating bottlenecks that delayed claim payouts to policyholders at their most vulnerable moments and posing both an operational and reputational risk.

QBE partnered with Hyperscience to deploy an intelligent document processing (IDP) platform combining NLP-driven AI with structured human-in-the-loop review. The platform extracts, classifies, and validates data from unstructured claim forms and invoices that vary significantly in format across QBE's global book of business, automating high-confidence extractions while routing low-confidence cases to human reviewers. The solution was integrated into QBE's existing claims workflows without overhauling core systems.

The implementation delivered an 80% reduction in claim processing time, with documents now processed in just 20% of the time previously required, a result confirmed by Jim Kinzie, QBE's Global Claims Transformation Lead. Document throughput on claim forms and invoices increased 5x, enabling faster claim payouts to customers experiencing adversity while the human-AI hybrid model preserved accuracy and oversight.

• Human-in-the-loop design is not a compromise: routing uncertain extractions to reviewers while automating high-confidence cases preserves accuracy at scale and eases staff adoption.
• IDP delivers outsized ROI in document-heavy workflows where volume is high and format variability is manageable with modern NLP.
• Prioritizing integration with existing systems over greenfield builds let QBE capture speed gains without replacing core claims infrastructure.