The IT team planned to migrate from this heritage system to a new, cloud-based ERP solution. Yet it needed to do so without losing any existing functionality of what had become the company’s communications backbone. More than 500 users across the organization’s headquarters, dealer network and subsidiaries relied on the software to perform functions for sales, spare parts management and post-sales assistance.

To enable a smooth transition, the team sought to improve key business processes before integrating them into the new system. One of these processes was the service entry process, which contains several diagnostic checkpoints that dealers must follow when performing car maintenance. The process held its own set of workflow challenges that if not addressed threatened migration success. Namely, several ERP system users followed process steps that had become common but undocumented knowledge.

To analyze its service entry process, the company selected the IBM® Process Mining solution, which applies data-mining algorithms to automatically discover process flows. It also simulates future processes from as-is processes. The Process Mining solution is included as a foundational capability across all IBM Cloud Paks for Automation offerings, including the IBM Cloud Pak® for Business Automation offering.

“We had already established a strong business relationship with IBM, but from a technical perspective the Process Mining solution offered capabilities that other solutions couldn’t,” says the ICT Process Manager. “The IBM solution enabled us to easily and effectively streamline the gap analysis process using data from our traditional ERP system and then perform what-if analyses. We could also quickly deploy the solution from the cloud.”

As soon as the company loaded service entry process data into the solution, an as-is process model, including all process variants, was automatically generated from the data. The undocumented steps were finally captured and recorded in the process model to keep them from becoming siloed information or forgotten steps.

The IT team next performed a gap analysis to get a visual comparison of the data-derived process model and the customer’s uploaded reference model. By comparing the reference model to the actual process, the customer immediately noticed a compliance issue in scan-in/scan-out activities. During the scan-in activity, workers connect the car to a diagnostic tool. They should also perform the scan-out activity, a best practice in which they check for warning indicators on the car’s control unit. The analysis showed that in 15% of the cases, workers did not perform the scan-out activity.

Additionally, by using the Process Mining solution, the IT team found that in about one-third of 75,000 cases, some dealerships abroad had opened the service entry case after the work start activity instead of before, resulting in nonconformance.