Cases › Usage-Based Insurance App Increased Customer Flow by 30%
Usage-Based Insurance App Increased Customer Flow by 30%
An app for an insurance company that helps clients acquire safe driving habits. App users get discounts for sticking to motor vehicle traffic regulations.
Goal
An auto insurance company requested our company to create a big data solution and an IOS/Android mobile app development with the purpose to:
- • monitor driving behavior
- • analyze collected data
- • score results and turn them into discounts.
Challenges
To release a high-quality product, our team had to optimize several processes:
- • Create the platform for administrators, agents, and support teams;
- • Design a user-friendly interface for drivers who sign up into the system;
- • Store and process large data scopes;
- • Optimize mobile phone battery consumption while tracking trips.
Solution
How we built it.
A multi-module system, each component independently deployable in Docker on Kubernetes, communicating via RabbitMQ or GraphQL.
Mobile App
We've kicked off the project with mobile app development for tracking trips. Our skilled mobile engineers have used, tried, and tested tools to customize every car trip. We ran numerous tests to optimize battery consumption and make the app distinguish car trips from rides on other vehicles. We've found a balance between motion tracking quality and battery drain. As a result, our developers designed a separate GPS tracking module to monitor mobile phone usage during the trip, affecting the outcome.
Data Stream Processing
Our engineers built up an IoT platform that obtains data on location and speed from mobile apps. It allows the platform to receive JSON data packages. Due to this information, the system can process, analyze, and plan trips.
Scoring analyzer
Our developers have created a system to calculate scores based on trip data. We used an ML model to assess and build up new trips in line with previous inputs.
Backend
Backend is a separate module that includes the business logic for registration in the mobile app, adding new vehicles, and receiving ready trips from the Data Stream Processing module. It also allows statistics to be processed and create email chains, SMS, and push notifications. In addition, the system has GraphQL API for the platform and mobile app management.
Management platform:
The management platform is the front-end app that lets administrators and insurance agents monitor the registration of new drivers and add vehicles to the system. Moreover, it allows editing email, SMS, and push notification templates sent to the end-users. It processes data obtained from the GraphQL client to manage all the users and generate different reports.
Infrastructure
All the modules are autonomous and set in a separate docker in Kubernetes. They interact with each other across RabbitMQ or GraphQL API. We use Kibana and an autoscaler to monitor and log data.
Platform features
What we shipped.
Driver/Vehicle
That mobile app screen shows the latest info about the driver/vehicle. It includes a map with the vehicle's current location, a part of the trip route if the vehicle is currently on the trip, status like 'Driving' or 'Idle', and info about the last trip (start/stop time, distance, duration, etc.), and an overview of Score & Discount.
Score & Discount
This mobile app screen shows the aggregated safe-driving scores for a given driver. It includes your estimated discount at the end of the program, program timeline (how many days of the program have passed), and driver's speeding/harsh braking info.
Tech stack
What it runs on.
Outcomes
Outcomes and Benefits
The launch increased the client's customer flow by 30%. The application set market trends for other insurance companies in usage-based insurance solutions.
Tell us what you're building.
40+ shipped engagements behind us. We don't take on every opportunity — but the ones where the problem fits the work we do exceptionally, we're glad to partner on, often for years.
Recognized by Clutch
HubSpot form (portal 242984681, form ba6582dd-2a5a-4698-8d78-3ddcc471cb4e) — loads on the client.
We work with a deliberate roster of clients each year. If we're not the right fit, we'll tell you — and point you somewhere that is.