Success Story: Same Damage, Different Angles – How a Pattern of Abuse Established through Repeated Photographs Was Revealed

In one of our large-scale fleet customers, claim files filed repeatedly in a short time began to attract attention. The files appeared independent of each other; different dates, different drivers and separate photos presented for each…

At first glance, each file was internally consistent. However, as the total cost increased, suspicions arose that there might be a different story behind this repetition.

This is where FraudFlow came into play.

First appearance: Files that appear to be independent damages

The process reviewed included:

• Multiple claim files opened for vehicles belonging to the same fleet
• Damage photos taken from different angles in each file
• Similar damage definitions based on parts
• Gradually increasing total payment amount

In traditional review, these files could be treated as separate events and taken into the payment process.

What did FraudFlow do? Read the memory of the image

FraudFlow took an in-depth look at the process thanks to its analysis engine that works not only with textual data but also with visual data:

• Visual similarity analysis: Uploaded photos compared at pixel and pattern level
• Damage trace matching: Locations of scratches, dents and broken areas were analyzed by superimposing them
• Angle normalization: Images taken from different angles were normalized to reveal traces of the same damage.

The result was striking: It was determined that the photographs presented in different files were actually versions of the same damage taken from different angles.

Red flags

During the analysis, the following critical signals emerged:

• Repeated damage areas on the same vehicle
• Even though the photos were uploaded on different dates, they have high visual similarity.
• Repeat damage descriptions with minor variations
• Files opened at short intervals for the same tool

These indicators pointed to a systematic pattern of abuse.

Operational challenge: The limit of the human eye

While there are hundreds of vehicles and thousands of damage records in fleet management:

• It is almost impossible to manually notice when the same damage is shot from different angles
• Cross-comparing between images requires significant time and attention
• Since operations teams generally make file-based evaluations, such repetitions may be overlooked.

FraudFlow eliminates this blind spot.

Result and impact: Multiple payment attempts from a single claim blocked

As a result of FraudFlow analysis:

• It was revealed that the same damage was filed repeatedly and payment was requested.
• Relevant files were linked and prioritized for review
• Unfair payment process stopped

In this way, the customer prevented unnecessary and repetitive payments, achieved serious control over fleet claims costs and increased operational efficiency.