Category | Activity | Example |
Pattern detection | Run continuous queries to detect and collect identified events. | Monitor Twitter feeds to detect and collect negative tweets that contain keywords like the name of your bank and “unhappy,” “slow,” or “bad.” |
Visualization | View the collected data in various formats. | Text of each complaint, a Google map of the complaint locations, and a graph of the process cycle times. |
Historical process mining | Examine historical data to determine whether similar complaints occurred in the past. | Aggregate fast process cycle times, aggregate slow process cycle times, and compare the results. Filter data to find the age of applicants in the slow processes. |
Root cause analysis | Drill into historical data to find the cause of the problem. | Applications from people under 18 in certain countries loop back through the process because of a form that requires a guardian’s signature. |
Process improvement | Improve the business process model. | Add a line to the application for a guardian’s signature. |
Real-time baselines, metrics, and alerts | Monitor key performance indicators (KPIs) and create rules that send alerts when data is out of compliance. | Set up a KPI for process cycle time and create a rule that alerts when cycle time is one standard deviation slower than normal. |