Virtual Agent Dynamic Intent Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Virtual agents are limited by pre-defined tasks and services, restricting user interaction to only those defined by the administrator, lacking the ability to autonomously adapt or understand user intent in real-time.
Innovation Solution
A virtual agent that analyzes user information and metadata to dynamically define custom intents at runtime, using machine learning processes to interpret user commands and execute actions based on intent services, dialogue engines, and session providers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If tasks and services are pre-defined by an administrator, then the virtual agent can execute specific functions, but the user is restricted to only those pre-defined tasks and cannot perform custom actions
Solution Approach 1:
The patent implements dynamic intent generation where the virtual agent creates custom intents at runtime based on user input and application context, rather than relying solely on pre-defined static intents. This allows the system to adapt to user needs dynamically while maintaining a structured foundation of pre-defined tasks.
Solution Approach 2:
The virtual agent autonomously generates custom intents by analyzing user input and application metadata without requiring administrator intervention. The system serves itself by automatically creating and executing custom tasks based on runtime conditions, eliminating the need for manual configuration of every possible user action.
2Ease of operation
If the virtual agent uses pre-defined tasks only, then the system structure remains simple and manageable, but the user experience is limited and rigid
Solution Approach 1:
The patent segments the intent processing into two distinct layers: pre-defined intents for common tasks and dynamically generated custom intents for user-specific needs. This segmentation allows the system to maintain simplicity for standard operations while providing flexibility for custom actions without overwhelming complexity.
Solution Approach 2:
The patent introduces an intermediary layer that analyzes user input and application metadata to generate custom intents. This intermediary process acts as a bridge between the simple pre-defined task structure and the need for flexible user interaction, translating user requests into executable actions without requiring direct modification of the core system structure.
3Adaptability or versatility
If the virtual agent dynamically generates custom intents at runtime, then users can perform a wide range of custom tasks, but the system requires complex machine learning processes and runtime analysis
Solution Approach 1:
The patent performs preliminary analysis by collecting and analyzing application metadata and user context before runtime intent generation. This preliminary action prepares the system with necessary information about user preferences, application state, and contextual data, enabling more efficient and accurate custom intent generation during actual user interactions.
Solution Approach 2:
The system implements feedback mechanisms where the virtual agent analyzes user responses and interaction outcomes to refine and improve custom intent generation over time. This feedback loop allows the system to learn from past interactions, improving the accuracy and relevance of dynamically generated intents while reducing the need for excessive runtime processing.
Data Source
AI summary
The present disclosure is directed techniques for executing a task or service using a virtual agent. A method includes: defining a plurality of intents; conducting a first tier of machine learning analysis to compare a received input string with a first subset of training phrases associated with the plurality of intents to extract one or more parameters of the received input string; conducting a second tier of machine learning analysis to compare an output of the first tier of machine learning analysis with a second subset of training phrases associated with the plurality of intents, wherein the comparison is used to generate respective similarity scores indicating whether the received input string matches one or more of the second subset of training phrases; selecting an intent from among the plurality of intents based on the respective similarity scores; and executing an action associated with the selected intent.