Taxi rides validation automation with human feedback

We automated a data-validation-before-billing system for our client. What previously required manual effort is now automated, allowing the client’s team to focus on reviewing critical data rather than combing through thousands of rows, making the whole process much more efficient. The automation ensures accuracy and consistency, reducing errors while still allowing human oversight for edge cases.

For our client, this means more timely and reliable reporting, enabling them to make informed decisions faster. Our solution also scales with the business, so they’re ready for future growth.

Let’s dive into this!

1. The Challenge: Improving a Manual Reporting Process

One of our clients in the taxi industry manages rides for corporate customers and needs to produce detailed reports for their top client. Their pre-existing process - while functional - relied on manual exports, Excel-based macros and manual hovering over thousands of lines to identify and resolve data inconsistencies as well as detect fraud. Though this system had served them for years, the increasing complexity and volume of data meant that it had become time-consuming to maintain and manually correct issues, using precious time from key resources. As the business grew, so did the need for a more efficient solution.

2. The Approach: Focused & Limited Development

Our goal was to move quickly and free the client’s key persons’ time, without going through the whole development process of a complete solution from scratch. That is why the system we designed used existing building blocks where possible, reducing the needs for custom development and speeding up the process. 

The solution we designed contains a dedicated data pipeline and an intuitive user interface, both enabling data labellisation, classification and correction, and offering a smooth collaboration between automation and manual review. This allows people to concentrate on the data needing a human eye only.

3. The Solution: Automating Data Flow and Reporting

We partnered with the client to design a custom data pipeline that enhanced and automated their reporting process. Here’s how the new system works:

  • Automated Data Extraction: We developed a process to automatically pull ride data from their platform, removing the need for manual exports and streamlining the first step.
  • Smart Issue Detection and Resolution: The pipeline now automatically identifies and resolves common issues. 90% of the rides are then not flagged for human verification.
  • Flagging for Manual Review: For more complex cases that require human decision, the system flags potential issues, allowing for quick and focused manual checks.
  • Efficient Report Generation: After the data is cleaned and verified, the system automatically generates the required report for the corporate client—quickly, accurately, and ready for delivery.

4. The Benefits: A Faster To Use System That More People Can Use

What previously required manual effort is now automated, allowing the client’s team to focus on reviewing critical data rather than combing through thousands of rows. The automation ensures accuracy as well as consistency and reduces errors, while allowing human oversight for pre-identified edge cases.

Effective testbench monitoring: Never lose a Simulink log again