.png)
Insurance
GEV uncovers image-based insurance fraud of at least 1.4%
with VAARHAFT
The Grundeigentümer-Versicherung uses VAARHAFT's image analysis API to quickly and automatically detect fake and manipulated claim images with AI-supported fraud detection.
✓
1,4% fraud rate for fake images
✓
min. 5-10€ savings per claim
✓
Fraud with AI-generated and AI-edited images is real
95,4%
Accuracy
AI build
in Germany
GDPR-compliant & No training with customer data
.jpg)
THE CUSTOMER
About Grundeigentümer-Versicherung VVaG
The long-established Owner's Insurance in Hamburg is one of the reliable partners for property owners in the Hanseatic city. Known for its tailor-made insurance solutions, it not only provides comprehensive protection against unforeseen damages but also plays a crucial role in preserving the value and safety of properties. With a broad portfolio of individually customizable policies and a strong commitment to service quality, the company ensures fast and transparent claims handling. Thanks to state-of-the-art technology and years of expertise, property owners in Hamburg can rely on completely worry-free and future-oriented coverage.
THE CHALLENGE
Detecting Fake Damage Images in Automated Claims Processing
For Grundeigentümer Versicherung (GEV), the detection of manipulated or fake damage images in insurance claims has become an increasingly urgent concern. As digital image submission becomes the norm in modern claims processing, fraudsters have found new opportunities to exploit the system. With the rise of advanced and easily accessible AI-based editing tools, even individuals with minimal technical skills can convincingly alter images to support false claims. This dramatic reduction in the barrier to entry for committing fraud poses a significant risk, both financially and reputationally, to insurance providers like GEV.
Traditionally, the company has relied on the trained eyes of claims processing staff to identify suspicious images. However, this manual approach is no longer sufficient. Not only has the volume of claims grown, but the sophistication of image manipulation has also advanced to a level where visual inspection alone often fails. Without the support of specialized detection tools like VAARHAFT, fraudulent claims could go unnoticed, slipping through the cracks and leading to substantial financial losses.
Compounding the issue is the fact that the frequency of image-based fraud is likely to increase, given how easy it has become to produce convincing fakes. This adds further pressure on GEV to find an efficient, scalable solution that not only flags manipulated images with high accuracy but also integrates seamlessly into their automated claims processing systems. Striking the right balance between robust fraud detection and streamlined operations has therefore become a central challenge in the company’s fight against digital insurance fraud.
THE SOLUTION
Automated AI-based image analysis for fraud detection and efficient claims processing
To proactively combat the growing threat of image-based insurance fraud, Grundeigentümer Versicherung (GEV) partnered with VAARHAFT in a joint pilot project that ran from September 2024 to February 2025. The primary objectives of this collaboration were to determine the proportion of manipulated images submitted as part of claims and to identify which insurance segments were most commonly targeted. Beyond that, the project aimed to assess the effectiveness of VAARHAFT’s image analysis API in identifying forgeries while supporting efficient and automated claims processing.
As part of the pilot project, GEV integrated VAARHAFT’s image analysis API directly into their claims processing systems. This integration enables the automated screening of all submitted damage images—regardless of whether they arrive via email, scans, or other channels. Within seconds of submission, each image is evaluated for authenticity, and the results are presented directly within the claim report, giving the claims team immediate insight into the image’s trustworthiness.
This automation significantly streamlines the overall workflow. Instead of manually examining images for signs of manipulation, claims processors now receive a consolidated damage report that includes VAARHAFT’s assessment alongside other relevant claim data. This not only saves considerable time but also reduces the risk of human oversight. Moreover, the API supports all common image formats, ensuring broad compatibility with the diverse types of media typically received in insurance claims.
While the system delivers a fast and reliable preliminary assessment, it functions as an intelligent early warning mechanism. Images flagged as potentially manipulated can be escalated for further investigation by experts, ensuring that automation does not compromise the accuracy and integrity of fraud detection. This careful balance between speed and scrutiny allows GEV to handle a rising volume of claims without sacrificing quality or due diligence.
THE RESULTS
1,4% Fraud rate
Of the 1.168 claims analyzed in the analysis period, 16 claims were subsequently confirmed as cases of fraud.
95,4% Accuracy
Only 4.6% of the analyzed claims were falsely marked as fraud. In all other cases, the solution was very precise and accurate.
> 100.000€ Savings potential per year
The savings potential is calculated on the basis of approx. 18,000 analyzed claims per year. With actual savings of 6.284€ for fraud cases in the analysis period, we realized a saving of min. 5-10€ per claim case.
During the analysis period from December 13, 2024, to February 28, 2025, a total of 1,168 insurance claims comprising 3,972 submitted images were examined using VAARHAFT’s image analysis technology. Of these, 16 claims were confirmed to involve fraudulent activity due to image manipulation, resulting in a fraud detection rate of 1.4%. The potential financial savings from identifying these cases during the pilot period amounted to €6,284. When extrapolated to the total number of claims GEV handles annually, this translates to an estimated savings potential of at least €100,000 per year through the continued use of VAARHAFT’s solution.
It’s important to note that the pilot project focused exclusively on image analysis. However, VAARHAFT’s capabilities go beyond just analyzing images—the system is also able to assess documents for signs of manipulation or forgery. This opens the door to uncovering additional forms of fraud that may currently go undetected.
Furthermore, the integration of VAARHAFT’s SafeCam offers an additional opportunity for reducing process-related costs. With the ability to validate the damage images marked as conspicuous in the downstream step with the secured camera web app in real time, SafeCam not only improves fraud prevention, but also streamlines the verification process and thus contributes to greater overall efficiency.
In summary, the results of the pilot project not only confirm the technical effectiveness of VAARHAFT’s fraud detection capabilities but also highlight its potential for significant long-term cost savings and broader applications in fraud prevention across multiple media types.
DOWNLOAD
Download the full story here!
Get the story as a summarized PDF-Document
.webp)