Half the claims never reach a human

 

The RFP was really a question about human intervention

SBI General, a quasi-public sector insurance company in India, issued an RFP in 2020 to modernise a claims processing workflow system built by a third party, and to introduce an AI-powered visual inspection process for vehicles arriving as claims. The stated aim was to maximise straight-through processing. STP is an unforgiving metric. It gives no credit to a process that is 90% automated; a claim either completes without human intervention or it does not.

That put the constraint on integration, not just on the AI.

The model had to work inside somebody else’s system. A damage assessment that produces an excellent report which then needs manual re-keying into a third-party estimation tool has not raised the STP rate by a single point.

The solution: AI assessment inside a third-party workflow

CamCom delivered an AI-powered remote automobile inspection system to SBI General’s surveyors, made available through a third-party app on their mobile phones.

• Guided image capture: Stencils and on-screen instructions ensure the surveyor captures the right images of the vehicle, first time.
• Full exterior assessment: Advanced deep learning methods assess damage across the vehicle’s exterior panels.
• A decision-grade report: Output covers the intensity of the damage, the area of the damage, and a repair- versus-replacement decision.
• Native integration: The assessment integrates directly with estimation data provided by a third party, so the claim moves through the existing workflow without a manual handover.

The results

• 50%-plus straight-through processing on claims, removing survey cost from more than half of all claim volume.
• Business continuity secured by eliminating human intervention from the assessment step.
• Significant survey cost savings, realised as a direct function of the STP rate rather than as a projected efficiency.

Engagement

3.5 years and ongoing. CamCom is the sub-contractor.