Customer-centric telematics solution

  • SparkT is our proprietary telematics solution focused on providing customer-oriented software for insurance companies that want to add the Pay-How-You-Drive model to their offer.
  • Our model combines multiple-source data to precisely evaluate every trip and provide drivers with contextual feedback.
  • The application is modular and white-labeled so it can be modified for seamless deployment for commercial clients.
  • The evaluation features are based on complex algorithms that process raw data to identify hazardous road behavior. The system is being supported by our ML-based extension - SparkSense.
cs-7.svg
A couple of years ago, our board members visited Japan as a part of CEE tech companies' representation in an international technology fair in Tokyo.

They were invited to give a presentation about telematics in a more intimate environment for a selected group of Japanese entrepreneurs. It turned out there was a company interested in our services.

The company was obliged to deliver a telematics solution to their client within three months. They were in trouble, as their contractor hadn’t delivered anything yet. The timeframe was tight, and our presence must have looked to them like the last glimmer of hope. We signed a contract to build an entire telematics solution from scratch in 3 months, keeping the project as our internal initiative, but giving the client a lifetime license. We developed and deployed the product on time for the client, which was a huge success. That gave us a strong foundation for building our own telematics system, which was extended and upgraded through the years.

UBI done right

We wanted to build a UBI (Usage-Based Insurance) platform to allow profiling drivers and setting individual rates based on their evaluation results. Insurance companies could acquire more low-risk drivers, increase UBI programs profitability, and improve road safety, as telematics solutions are proven to positively influence end-users' driving habits.

Most UBI programs incorporated the Pay-As-You-Drive model in recent years, but it only took time and distance factors for the evaluation. Our goal was to enrich it with additional data to align with a more up-to-date, customer-centric model - Pay-How-You-Drive. This way, we could address driving habits more precisely by giving contextual information about any trip, especially considering hazardous events.

The system performs a set of operations that combine to our new model:

  • Acquiring information about the trip, such as GPS points, speeds, acceleration and braking data, time, gathered from various sensors from the smartphone and IoT devices;
  • Enriching collected data with contextual information: road types, speed limits, real-time traffic or weather conditions;
  • Identifying hazardous events such as rapid accelerations and brakings, speedings, harsh cornerings;
  • Processing the data by a set of algorithms to profile driver and evaluate the trip in terms of safety and economy;
  • Visualizing the data and providing the driver with feedback and improvement tips;
CCTS-screens.svg
A well-oiled machine

The architecture design part of the development cycle was crucial. We wanted our solution to be universal but personalizable to suit every client who wanted to incorporate it. The other important factor was ensuring the system’s scalability, as its proper functioning required continuous transfer of lots of data from multiple sources.

We decided to follow a modular approach, dividing the system into separate, deployable parts. This way, we could address new clients who didn't want to proceed with the full-featured version and companies already using telematics but wanting to extend its functionalities.

The last part of the design focused on keeping the application white-labeled, so that every company could customize UI elements with their key visuals. That allowed the application to serve as a marketing channel, increasing the end user's brand awareness, loyalty, and engagement.

Research and development

As much as we knew about telematics, we had to propose and test many ideas during the development. The most important ones were:

  • Working out the road event detection algorithms, as we had to examine which data generated by sensors will be useful and what frequency of readings will guarantee a satisfying level of precision in assessments,
  • Defining and developing driving evaluation algorithms to know which factors should we consider and how to pair them with driving economics and safety indicators
  • Optimizing data transfer, processing, and storage to keep the cost low without compromising performance metrics
  • Operationalizing smartphone-based telematics by allowing the application running in the background to generate data with high frequency without draining the battery
  • Replacing commercial cartographic data providers with our proprietary, open data-based solution to drastically reduce operational costs
Summary

Our final solution is a fully-featured, modular, white-labeled telematics system. It gathers information from the driver's phone, optional BLE (Bluetooth low energy device), and OBD2 (on-board diagnostics) devices installed in the vehicle, enriching it with 3rd party data to provide a broad assessment of the trip and driver's behavior.

The latest addition to the system is an innovative, intelligent dashboard camera powered by a set of ML algorithms. You can read more about it here.