Technology
Technology OverviewImage Trust LayerDocument Trust LayerAudio Trust LayerInformation Trust LayerSource VerificationProducts
Fraud ScannerSafeCamSafeMicGardioIndustries
InsuranceBanking & KYCAuthoritiesMedia & PressHR & ExpensesDefenseResources
BlogCase StudiesAPI DocsPress AreaCompany
About UsJobsContactManipulated claims images and forged documents cost the industry billions. VAARHAFT gives you the tools to automatically verify every submitted piece of evidence, and actively resolve dubious cases. In seconds. Directly within your claims process.
Accuracy figures are based on internal benchmarks.
Trusted by
.webp&w=256&q=75&dpl=dpl_UCTwikveTDJ5bPjn2FpCxK3YLrqa){kind=small}
%20(1).webp&w=256&q=75&dpl=dpl_UCTwikveTDJ5bPjn2FpCxK3YLrqa){kind=small}
.webp&w=256&q=75&dpl=dpl_UCTwikveTDJ5bPjn2FpCxK3YLrqa){kind=small}
Any insurer digitizing claims processes without systematically verifying submitted media is opening the door to fraudsters.
Sources: GDV · Sumsub Identity Fraud Report 2023
For your claims management, this means:
Can you tell which damage is real?
Make your guess!
VAARHAFT bridges the gap between fraud detection and fraud prevention. Our solution consists of two components that are powerful individually, but only reach their full potential when combined:
The Fraud Scanner is VAARHAFT's AI detection module. It automatically checks every submitted image and document for visual authenticity, fully automated, in under 4 seconds, directly through your existing claims software.
Available as a REST API for automated integration into your processes and as a web tool for manual ad-hoc review by claims adjusters.
When the Fraud Scanner detects an anomaly, the policyholder automatically receives an SMS with a link to the secure camera web app, no download required. The customer photographs the damage again, live and in real time.
Honest customers complete the process in under one minute. Fraudsters are blocked during the capture process and frequently withdraw their claim voluntarily.
The Fraud Scanner detects fraud. SafeCam proves it, or clears the suspicion. Together, they drive your false positive rate to near zero.
Every submitted image and document undergoes a comprehensive forensic analysis in parallel. You receive a structured result with confidence scores, explanations, and actionable recommendations.
Detects AI-generated damage photos with up to 99% accuracy before they enter your claims process.
Visualizes exactly where and how an image was manipulated. Transparent and comprehensible for claims adjusters.
Instantly identifies if a submitted damage photo already exists on the internet, a reliable indicator of fraud.
Automatically checks whether photo dates match the reported damage date. In this case, a 12-year discrepancy.
Detects AI-generated invoices and documents and checks contact details for plausibility. Here: 98% AI probability and an invalid IBAN.
Reveals capture location, device information, and timestamps for automated plausibility checks.
Fraud Scanner and SafeCam intervene at different points in your processes, from post-submission image verification to secure on-site documentation and active resolution of suspicious cases.
A policyholder reports a water damage claim and submits photos through the claims portal. The images look plausible at first glance, but the Fraud Scanner detects within seconds that the damage was retroactively inserted into an authentic photo using AI. The heatmap marks the exact manipulated areas. Additionally, the duplicate check reveals that the same base image was already used in a previous case. The claim is automatically flagged as RED and presented to the claims adjuster with the full analysis report, before any settlement is made.
Ready to see it in action?
Book a DemoEvery prevented fraud case improves your loss ratio. With automated detection at the point of claim submission, you reduce fraud losses before they hit your books, directly strengthening your profitability.
Automated AI analysis replaces expensive manual review processes. Your fraud team focuses only on flagged cases, reducing operational costs while increasing detection rates.
Proprietary deep learning models trained on our own datasets. Weekly updates keep detection at the cutting edge.
By combining Fraud Scanner and SafeCam, suspicious cases are not just detected but actively resolved.
No permanent storage of customer data. Images deleted immediately after analysis. Hosting on AWS Frankfurt.
The SafeCam camera web app is billed purely on a success basis: you only pay when fraud is identified and prevented.
The SafeCam camera web app is fully adaptable to your corporate design: logo, colors, texts, and communication flow.
Continuous monitoring of new threats. Automatic model retraining. Updates rolled out server-side with zero IT effort.
Two ways into your claims processes: deeply integrated via API or ready to use in the browser. Both run on the same analysis engine.
Integrate the VAARHAFT REST API into your claims management system. Every submitted image and document is analyzed automatically at claim intake, ideal for straight-through processing.
Start without any technical integration: claims handlers upload suspicious claim images directly in the browser and receive the full analysis report in seconds.
Technical Specifications
ITC DIA Amsterdam
2024
Versicherungsforen Leipzig
2024
Insurtech Hub Munich
2024
insureNXT · InsurLab Germany
2024
Versicherungsforen Leipzig
2023
Typical integration takes 1–4 project days. The web tool can be used immediately without any technical integration.
No. Images are deleted immediately after analysis (dwell time: 2–3 seconds). No customer data is stored and no customer data is used to train our models.
Overall accuracy is above 95%, measured against internal representative datasets. Models are retrained weekly with current data.
In principle, yes. However, the solution only reaches its full effectiveness when both work together: the Fraud Scanner identifies suspicious cases, SafeCam resolves and deters them.
Exclusively on Amazon AWS in Frankfurt, Germany. All communication is encrypted via HTTPS.
JPG, JPEG, PNG, TIFF, HEIC, and PDF. For PDF files containing embedded photos, these are automatically extracted and analyzed individually.
You only pay for SafeCam usage when fraud is actually identified and prevented. The Fraud Scanner API is billed separately based on analysis volume.
Yes. The camera web app is fully available as a white-label solution: logo, colors, texts, and the entire communication flow are individually configurable.
See in a 45-minute live demo how VAARHAFT integrates into your claims processes, and how quickly you'll achieve your first results.
