RPA Task Variant Detection via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robotic process automation (RPA) technologies lack the ability to discover variants of automatable tasks effectively, as they are either passive or require manual user input, failing to capture and analyze user interaction data efficiently.
Innovation Solution
The system generates task flow data using a combination of task mining and user input, employing machine learning models to identify user interaction data from screenshots, paths, and actions, enabling the determination of variants of automatable tasks and associated metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If task mining is used to automatically discover automatable tasks, then productivity is improved, but the ability to detect variants of tasks deteriorates
Solution Approach 1:
The patent combines task mining and task capture approaches into a unified system. Task mining automatically records user interactions to generate task flow data, while task capture allows users to manually record and annotate task variants. By merging these two methods, the system achieves both high productivity through automation and comprehensive task variant detection through user input.
Solution Approach 2:
The system performs preliminary task recording and analysis to build a comprehensive task flow database before automation deployment. By预先 collecting task variants and user interactions, the system can better identify and handle different task scenarios, improving both automation efficiency and adaptability to task variations.
2Measurement precision
If task capture is used with manual user input, then task recording accuracy is improved, but productivity deteriorates
Solution Approach 1:
The system applies task capture selectively only for tasks that require variant detection or complex user input, while using automated task mining for routine tasks. This partial application of manual recording maintains accuracy where needed without sacrificing overall productivity.
Solution Approach 2:
The system enables users to self-record and self-annotate task variants through the task capture interface, reducing the need for extensive manual analysis and processing while maintaining high recording accuracy.
3Ease of operation
If conventional task mining is used without user input, then ease of operation is improved, but the ability to capture additional content deteriorates
Solution Approach 1:
The system incorporates feedback mechanisms where users can review automatically detected tasks and provide corrections, annotations, and additional information. This feedback loop maintains ease of operation while capturing valuable user insights and task variants that automated mining alone would miss.
Solution Approach 2:
The system dynamically adjusts between automated task mining and user input collection based on task complexity and user preferences. For simple tasks, automated mining suffices; for complex tasks requiring variant detection, the system activates user input mechanisms, optimizing both usability and information capture.
Data Source
AI summary
Systems and methods are provided for determining variants of an automatable task. Task flow data of a performance of an automatable task by one or more users is received. The task flow data is generated based on user input using task mining. User interaction data is identified from the task flow data. One or more variants of the automatable task are determined based on the user interaction data using a machine learning based model. The one or more variants of the automatable task are output.