Case Study: Automated Car Diagnostic Tool

Overview

A large car dealership sought to improve their customer experience by resolving inefficiencies in the workshop and improving client-service agent communications when the client drops their car off for repairs. The amount of time taken by the diagnostic technicians to review the backlog of vehicles and their symptoms was affecting the way the garage was communicating with clients, leading to low satisfaction and inefficient work scheduling.

They sought an improved repair diagnosis approach that would:

  • Provide faster diagnosis allowing increased productivity for technicians.
  • Allow the service agent to discuss costs before the client leaves the dealership, increasing transparency.
  • Improve the success of cross-sell and up-sell opportunities while face to face with the client.
  • Enable more efficient and faster workload scheduling in the garage.

How did Heron approach this?

We analysed a range of structured and unstructured data to identify patterns and trends to be taken into consideration for the diagnosis. Through creating a conversational platform to capture additional information from customers on arrival, and modelling best-practice for diagnosis by mechanics in an automated decisioning model, our tool combines and analyses all of this information to provide a recommended diagnosis.

Data Up: Machine Learning was used to build a model of relationships between symptoms and faults based on data from multiple sources, which can be mapped to the required repairs and estimated costs. The tool includes an integrated feedback loop so that it can improve over time, and provides service agents and customers with confidence that the diagnosis is correct.

Human Down: Heron spent time with both Service Agents and mechanics to model their diagnostic approach in a knowledge map that would allow better questions to be asked of the client in order to provide more accurate diagnosis of the problems experienced.

The diagnosis then was presented with the estimated costs for the service and a predictive model capturing any near-future faults that might occur, to provide the customer with full visibility upfront. The result was a solution with an integrated feedback loop that provides both Service Agents and customers with the confidence that the diagnosis and recommended repair is correct and aligned to the pricing options provided.

 

Technologies Used

  • Machine Learning
  • NLP/NLU Chat interface
  • Automated decision making

The Outcome

We designed a system that provides both service agents and customers with the confidence that the diagnosis and recommended repair is correct, with the diagnosis aligned to the pricing options, timeframes and parts required. The integrated feedback loop provides both Service Agents and customers with the confidence that the diagnosis and recommended repair is correct and aligned to the pricing options provided.

The system is supported by predictive model to inform what the client can expect over next six months.

With faster, more accurate vehicle repair diagnostics, Heron enabled the car dealership to offer customers a quality of service that will keep them coming back.