HDFC ERGO achieves 98% straight-through processing in motor insurance with AI-driven underwriting and claims automation

India's general insurance sector processes millions of motor and health claims annually under intense pressure to reduce turnaround times while containing fraud losses — challenges that scale directly with portfolio size. HDFC ERGO General Insurance, one of India's largest private general insurers, found its traditional rule-based underwriting increasingly inadequate for dynamic, personalized risk pricing across diverse customer segments. Manual adjudication created bottlenecks in claims settlement, while reactive fraud detection meant losses were often identified only after payouts. Across both motor and health portfolios, the absence of zero-touch processing kept operational costs high and customer satisfaction constrained, with no systematic mechanism to feed claims outcomes back into underwriting decisions.

HDFC ERGO embedded AI and ML models across its full underwriting and claims lifecycle, replacing static rule engines with continuously learning systems. For underwriting, models ingest geospatial intelligence, historical claims records, vehicle telematics, and health parameters alongside external data signals — all connected in a closed-loop feedback architecture that refines underwriting rules as claims outcomes accumulate. Claims automation relies on computer vision for vehicle damage assessment and cost estimation directly from submitted images, eliminating manual inspection for standard cases. GenAI-powered agentic systems digitize and analyze third-party legal claims documents, while dedicated ML models proactively detect fraud by flagging anomalous patterns in claims frequency, provider behavior, and documentation. Contact centre operations were simultaneously transformed through GenAI-driven voice bots and chatbots, enabling 24×7 customer service without proportional headcount growth.

The program delivered near-total automation in motor insurance, with HDFC ERGO achieving approximately 98% straight-through processing (STP) on motor claims — meaning roughly 98 in every 100 motor claims are adjudicated and settled without human intervention. Health insurance, a structurally more complex portfolio with greater clinical variability, reached approximately 65% STP, a substantial reduction in manual workload. Key outcomes include:

• Faster settlement cycles across both lines through zero-touch adjudication on eligible claims
• Proactive fraud containment via ML models surfacing anomalies before payouts are released
• 24×7 contact centre coverage enabled by GenAI voice and chat agents without equivalent staffing increases
• Compounding underwriting precision as closed-loop feedback continuously sharpens risk models

• Differentiate by portfolio complexity: Motor insurance reached 98% STP faster than health because damage assessment is more structured — plan automation sequencing accordingly.
• Closed-loop architecture is the multiplier: Feeding claims outcomes back into underwriting rules compounded efficiency gains beyond what standalone models deliver.
• Human-in-the-loop is a governance requirement, not a limitation: Preserving underwriter oversight on complex cases maintains regulatory accountability and model transparency.
• Evaluate GenAI ROI enterprise-wide: Measuring impact only on individual use cases understates value — contact centre, fraud prevention, and underwriting gains must be assessed together.
• GenAI in legal claims is an underexplored lever: Automating third-party claim document analysis can unlock significant cycle-time reductions in contested or litigated cases.