Research &
Development 

at ControlExpert

Research & Development

As pioneers of digitalisation, we continually seek ways to broaden our digital frontier to integrate the very latest emerging technologies to further simplify and enhance our processes.

 

To enforce this philosophy, we established our own Research and Development team in 2015.

 

Our R&D team of 30 members has been intensively researching and creating our next generation of services. For example, how leveraging on-board sensor technology can be used for car damage notification, localisation, and calculation. Another example of our research is how image recognition using machine learning can help detect damage levels and parts requirements.

 

Our goal is to continue to improve our applications and products to provide our customers with the best tools needed to generate desirable results.

30 data scientists from the field of:

Telematics – Claim information in seconds

Vision:

"In the event of an accident, the damage can be quantified before the car even comes to a standstill."

  • Data acquisition by sensors already present in the vehicle.
  • Reliable information on the circumstances of the accident and the extent of the damage.
  • Augmenting this data with the CE database to automatically create an estimate.

Automated Image Recognition

"Active in the same field as Facebook, Google, and Apple, where they apply efforts into facial recognition, we at ControleExpert use the same methodology and algorithms to images of cars."

  • Inspection by automotive mechanics is made plausible and easier.
  • Images can be automatically assigned to damage zones.
  • In the future, an automatic pre-check will be possible.
  • Faster and more efficient processes.
1

Detection

Identify and classify parts from photos

2

Assessment

Identify the damage of the parts

3

Evaluation

Identify the degree of damage of each part

4

Calculation

Determine the expected repair costs

Chatbots & language assistants

“Chatbots or language assistants simplify the initial contact process with the customer. The claim handler is then given more time to focus on more complex or specialized customer requests."

  • Process acceleration.
  • 24 / 7 availability.
  • Increased customer satisfaction.

We are successful with Chatbots!

 

2016 – InsurHack (Zurich): 
„Claim settlement with Chatbots” (2nd)

 

2017 – HackNEXT (Allianz): 
„Carlexa“ (2nd)

3D Printing

“The replacement parts industry will be revolutionized. Development, manufacturing, and sales will need to be completely reconsidered.”

Research project “It’s digitive”
 
Supported by:

AI-based damage recognition
on photos and video sequences

"AI-based detection and calculation of vehicle damages on photos and video sequences"

Research project „Bergische Innovationsplattform für Künstliche Intelligenz“ (BIT) (Bergisch innovation platform for artificial intelligence)
 
  • Research question:
    Do newest model architectures learn better on photos or videos of damages?
  • Task ControlExpert:
    • Development of machine learning models for detection and assessment of damages on photos and videos
    • Testing scientific machine learning models on real-world-data in cooperation with universities
  • Participating partners:
    Bergische Universität Wuppertal, Hochschule Bochum, Institut für Qualitäts- und Zuverlässigkeitsmanagement GmbH, Lorent IT-Lösungen GmbH.
Supported by: