Know exactly what happened.
From suspicion to evidence in minutes. AI-powered claims reconstruction.
80% of suspected fraud gets paid — not because it didn't happen, but because no one gathered the evidence in time.
From suspicion to proof.
Automated Claims Reconstruction
Analyzes vehicle movement 72h before and after the event. Route reconstructed with GPS, LPR cameras, cell towers, and behavioral data. Timeline: where it was, speed, stops, interactions. Detects inconsistencies between declared version and objective data.
What manual investigation takes weeks, Lens delivers in minutes. No field expert needed to prove the vehicle wasn't where the insured claimed.
LPR/OCR Camera Network
Access to thousands of public and private license plate reader cameras. Historical query by plate and period. Correlation with theft and block records. Cloned vehicle identification. Photographic evidence with timestamp and location.
If the insured says the car was stolen Monday night in Campinas, but cameras show it driving Tuesday in São Paulo — that's evidence. Not suspicion.
Behavioral Fraud Detection
Cross-references historical behavior with the reported event. Suspicion score based on pre and post-event anomalies. Fraud patterns: arson, staged collision, simulated theft. Cross-check with claim history across insurers. Recidivism alerts.
Insurance fraud is sophisticated because fraudsters know the process. Lens uses behavioral patterns to detect what the official version hides.
Legal-Grade Evidence
Technical report with chain of custody. Timestamped, traceable evidence by source. Documentation for judicial and administrative proceedings. Solid basis for claim denial.
Denying a claim without evidence exposes the insurer to litigation. With Lens, the denial comes with a complete technical dossier.
Claims Workflow Integration
API integration with any claims management system. Auto-activation when suspicion score exceeds threshold. Claims dashboard with integrated investigation panel. Configurable SLA by claim type.
Manual process doesn't scale. Lens auto-triggers on suspicious claims so your team focuses on what matters.
In practice.
Simulated theft — vehicle appears on cameras 3 days later
Expert visits site 2 weeks later. Standard process. Claim likely paid.
→LPR cameras detect the vehicle 3 days after the alleged theft, near the insured's address. Reconstruction shows normal use. Fraud evidence in 48h. Claim denied with technical dossier.
Staged collision — two vehicles with incompatible history
Two vehicles report collision. Appear unrelated. Both claims paid.
→Historical movement shows both vehicles at the same location multiple times. Cross-check with Zarv ID: owners have indirect corporate link. Both denied.
Asset recovery — defaulted vehicle located via cameras
Slow, expensive manual search. High non-recovery rate.
→LPR cameras locate the vehicle in a specific neighborhood with recurring pattern. Recovery team dispatched with precise location. Recovery rate: 85–93% vs. 45% market average.
Proven results.
Lens vs. manual investigation
vs. 45% market average
Every denial with evidence reduces it forever
Only 20% of suspected claims are proven today
Proven at scale.
TagPro
Insurer · Auto Insurance (B2C)
New clients onboarded via Zarv ID (KYC + scoring)
Portfolio monitored in real time via Zarv Signal
Losses investigated via Zarv Lens with evidence generation
Arval / BNP Paribas
Leasing + Self-Insurance · Corporate Fleet (B2B)
Fleet monitored continuously via Zarv Signal
Claims reconstructed automatically via Zarv Lens
Zarv ID integration underway for new lessee onboarding
Ready to integrate.
By scope / per claim investigated. Value model: fraud avoided vs. investigation cost. Pilots with sample of historical claims.
Your data, protected by design.
See risk before it costs you.
GDPR & CCPA Compliant · No commitment · Integration in days