FlexyPro Micro Insurance Platform
Innovative AI platform for micro car insurance solutions.
AI-Powered Advanced Risk Modelling
At FlexyPro platform, our AI-powered risk model is built upon a deep analysis of over 1 million real-world trips from tens of thousands of users, combined with verified insurance claim data. To process this data efficiently, we developed a suite of automated tools for filtering, normalization, and enrichment. Using correlation and machine learning analysis, we identified several behavioral and contextual factors that significantly contribute to elevated insurance risk.
Driver Digital Model
In FlexyPro we implemented the Driver's Digital Model. The Driver’s Digital Model is a virtual profile of the driver built on data regarding their behavior, preferences, trip history, and risks associated with their driving style, weather conditions, road conditions, and other factors. This model not only allows for risk assessment to calculate insurance costs but also actively interacts with the user through triggered communications, offering personalized proposals and services.
Differentiation
What is the difference between voluntary comprehensive insurance for a short duration and FlexyPro?
Competitive Advantages
Risk Modelling
Very broad set of parameters for dividing users into groups
Аlexible additional training of the risk model, supplementing the risk model with additional parameters
Mobile Centric Solution
Market leading accuracy of event detection using the phone while driving
Large number of definable types mobility movement by car, taxy, scooter/bike, bus, subway, train, etc.
AI/ML Technologies
We are using the most applicable AI/ML technologies to increase data accuracy and to shorten time to market.
K‑d Tree (K‑dimension Tree) Visualisation
2D Space Partitioning Visualization with Tree Building and Search Zones


Nearest Neighbor Search in k‑d Tree


Support Vector Machine (SVM)
Illustration with hyperplane, margine zones and reference vectors


K-d Tree With Dense Mesh
Example of dense space partitioning, visualization of tree depth through lines of different color/thickness


Random Forest Claffifyer


Gradient Boosting
The picture shows how successive trees correct errors of previous ones (red/green dots reflect corrections)
Step-by-step construction scheme: tree - error - next tree - improvement.

