Visada

Computer vision platform for detecting fraud in P&C insurance claims

Claims Technology Startup

Overview

Visada is a fraud detection technology provider for property and casualty insurers, specializing in automated image analysis of vehicle damage claims. The company builds AI tools for P&C carriers that need to systematically review claim photo evidence for signs of fraud rather than relying on manual sampling.

The company's platform, DAFNE (Digital Anti-Fraud Neural Engine), applies proprietary computer vision models to vehicle accident images. It constructs an indexed archive from years of prior claims and then compares incoming claims against that database to surface reused photographs, pre-existing damage, and recurring visual patterns across claims filed at different points in time. A human-in-the-loop step has specialist analysts verify high-confidence matches before the system generates structured evidence reports, which are designed for internal anti-fraud workflows and formal proceedings.

Visada was co-founded by Angelo Capolupo, who brings over 20 years of insurance claims assessment experience, and Giuseppe Tortorelli, a computer science engineer from the Polytechnic University of Milan who leads platform architecture. The company is Italy-based and operates with support from EU co-financing.

Products & Services

DAFNE (Digital Anti-Fraud Neural Engine)

DAFNE is an AI platform that analyzes vehicle damage images from P&C insurance claims to identify fraud indicators. It processes historical claim archives to build a searchable index of damage patterns, then matches each new claim against that history to detect reused photos, pre-existing damage, and recurring visual inconsistencies. Each automated finding goes through analyst review before a final report is issued.

Key Features

Pixel-level vehicle damage analysis and classification of damaged areas
Historical claim archive matching to detect reused or recurring damage across time periods
Automatic anonymization of personally identifiable image content (license plates, faces)
Human-in-the-loop verification by anti-fraud specialists before reporting
Auditable evidence reports with side-by-side visual comparisons, suitable for internal use and litigation support

Target Users: Anti-fraud teams at P&C insurance carriers