Automated Knock-Out Checks with Process Mining

The text discusses overprocessing waste in business processes and highlights the use of knock-out checks to identify and eliminate this waste. A novel, interpretable process mining approach is proposed to detect improvement opportunities and recommend process redesigns. The paper presents an implemented software tool and showcases its applicability on real-life processes.

Knock-Out Checks and Overprocessing

Knock-out checks are tasks in a business process that classify cases as either "accepted" or "rejected". If a case is rejected, it triggers the execution of an anchor, a pre-determined point in the process, effectively ending the process. When a knock-out check leads to a case's rejection, all the work up to that point is deemed overprocessing waste, leading to inefficiency in terms of time and cost.

Knock-Out Check Discovery and Analysis

The proposed method uses an event log to discover and analyze knock-out checks. The process involves discovering knock-out checks, their decision rules, and dependencies from the event log, then assessing overprocessing waste and effort-per-rejection rates. It supports three modes of operation: semi-automatic discovery, automatic discovery, and known knock-out checks.

Improvement Opportunity Identification

The final step is to identify improvement opportunities and suggest possible redesigns based on the knock-out checks. These redesigns include knock-out reordering (least effort to reject), knock-out relocation (earliest possible point), and knock-out rule adjustment (changing attribute values).

Overall, the process mining approach provides insight into process redesign opportunities to reduce overprocessing waste. The interpretability of the results supports decision-making and increases confidence in redesign recommendations.