Intent-Based Task Mining for User Intent and Variant Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robotic process automation (RPA) tools lack the ability to determine user intent and identify task variants, focusing only on user actions rather than the underlying intent behind those actions.
Innovation Solution
Implementing intent-based automation using AI/ML models to analyze vector representations of user interactions, clustering and classifying task types, and mapping them to RPA workflows, with additional context provided by application programming interfaces (APIs) to understand user intent.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If task mining tools record user interactions and use AI/ML models to identify automatable tasks, then task identification capability is improved, but the tools still cannot determine user intent behind actions
Solution Approach 1:
The patent introduces an intent detection module as an intermediary between task recording and automation workflow generation. This module analyzes user actions within their contextual environment (application state, sequence of actions, temporal patterns) to infer underlying intent, thereby recovering the lost intent information without compromising task identification accuracy
Solution Approach 2:
The patent adds a new dimension of analysis by examining the temporal and contextual dimensions of user interactions. Instead of analyzing isolated actions, the system analyzes sequences of actions across multiple screens and applications, transforming the data from a single-dimension (action type) to multi-dimension (action sequence, contextual state, temporal pattern) to enable intent inference
2Ease of manufacture
If RPA tools focus on recording user actions without analyzing intent, then implementation simplicity is maintained, but automation effectiveness deteriorates
Solution Approach 1:
The patent implements self-service automation where the system automatically detects intents and generates appropriate automation workflows without requiring manual configuration. The intent detection module autonomously analyzes user behavior patterns and translates them into executable RPA workflows, maintaining implementation simplicity while dramatically improving automation effectiveness
Solution Approach 2:
The patent changes the fundamental parameter of automation from action-based to intent-based. By detecting the underlying intent behind user actions and using it as the basis for workflow generation, the system transforms the automation approach while keeping the implementation process simple and automated
3Quantity of substance
If task monitoring tools track time spent on applications, then usage analytics are provided, but the actual tasks being performed remain unidentified
Solution Approach 1:
The patent extracts meaningful task context information from the collected usage data by analyzing sequences of screen transitions, application states, and interaction patterns. Instead of treating data collection as an end in itself, the system extracts the essential task context that reveals what users are actually doing, converting raw data volume into actionable task intelligence
Data Source
AI summary
Intent-based automation that discovers automatable tasks and/or determines task variants in data is disclosed. Task capture data may be utilized to determine task variants in task mining data. Semantic understanding of user actions by artificial intelligence (AI)/machine learning (ML) model(s), for example, may be applied to determine the intent of the user rather than only focusing on what actions the user is performing on the computing system. Application logs and semantic understanding may be used to facilitate a more accurate determination of what the user actually intends to do. Task capture for individual user flows may be performed. Once these are captured, task capture algorithms and AI/ML models are used to determine which parts of the flows are similar and/or match and which parts are unique. The path through these flows can then be followed to build a process graph that includes decision points representing the unique flows.