Task Mining: User Interaction Data Collection
Accelerate operations and achieve process excellence by aggregating data from employee desktop tasks and turning them into understandable behavioral patterns. Gain a complete picture of the process by seeing how employees’ actions — on data entry screens, in email, spreadsheets, websites, and other tools — affect process outcomes.
Capabilities
How does Celonis Task Mining work?
Aggregate user interaction data
Celonis process visualization pinpoints the specific desktop actions and systems that are creating friction in the overall process and reveals which activities can improve outcomes.
Analyze patterns in desktop tasks
Celonis Task Mining aggregates noisy desktop actions and turns them into clear activity patterns, revealing which sets of actions create friction and which ones improve outcomes.
Align user desktop behavior to overall business outcomes
Celonis connects user behavior to system-level processes, so that you can reduce behaviors that results in negative outcomes and encourage the ones that drive positive outcomes.
Task Mining in Action
Identify time-consuming tasks, find where processes are getting stuck waiting for human intervention, and validate the complete and best set of tasks to achieve the best business outcomes.
Task Recorder
The Task Recorder gathers user activity on desktop machines by capturing user clicks, application actions, and screenshots. The Task Recorder is enabled with user privacy features that enable black listing of applications, hashing and obfuscation of sensitive data like passwords.
Optical Character Recognition (OCR)
After collecting the desktop data, Celonis turns that noisy unstructured data into words, by using OCR technology that is able to spot keywords, pull out specific case IDs, and provide a textual data set for mining.
Natural Language Processing and Clustering
Celonis NLP “reads” the textual data, and groups the data into “actions” and “activities” (such as researching a customer or updating an order) through clustering algorithms. These actions and activities provide a means to understand and analyze user behaviors.
Case Matching
The enriched and clustered user interaction log allows matching based on IDs, usernames, keywords, and labels, connecting the User Interaction Activities to the Business Event Data and thereby providing a complete view of the entire process: human activity and system activity.
