Virtual Agent System for Unstructured Conversations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current chatbots are limited to handling single defined tasks and struggle with unstructured conversations that involve multiple unlinked tasks, making them ineffective in resolving complex user requests and lacking reusability across different applications.
Innovation Solution
A virtual agent system that incorporates automatic speech recognition, natural language understanding, conversation management, and reusable conversation block sub-systems, allowing for the generation of predefined conversation maps and agent intents to handle complex user requests across various tasks without reconfiguration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If chatbots are designed to handle multiple different tasks, then the capability to handle unstructured conversations is improved, but the device complexity and lack of reusability across applications worsens
Solution Approach 1:
The chatbot system is segmented into independent, reusable task modules (e.g., booking module, complaint module, enquiry module) that can be selectively activated. Each module handles a specific task independently, allowing the system to manage unstructured conversations involving multiple tasks without creating a monolithic complex design. This modular segmentation enables reusability across different applications while maintaining the capability to handle diverse unstructured inputs.
Solution Approach 2:
The patent implements a universal chatbot architecture with a common conversation management layer and intent recognition system that can orchestrate multiple specialized task modules. This universal framework allows the same base system to be deployed across different applications by simply configuring different combinations of task modules, achieving both multi-functionality and reusability without redesigning the entire system for each application.
2Device complexity
If chatbots are designed for single defined tasks, then the device complexity is reduced and reusability is improved, but the capability to handle unstructured conversations with multiple tasks worsens
Solution Approach 1:
The patent merges multiple simple, single-task chatbot modules into a unified conversation system. Each module maintains its simplicity for handling a specific task, but they are combined through a central conversation management layer that can sequence and coordinate multiple tasks within a single conversation flow. This merging allows the system to handle unstructured conversations with multiple unlinked tasks while preserving the simplicity and reusability of individual task modules.
3Manufacturing precision
If chatbots are specifically designed for certain applications, then the manufacturing precision for that application is improved, but the adaptability to other applications worsens
Solution Approach 1:
The patent creates reusable task module templates that can be copied and configured for different applications. Each task module (e.g., booking, complaint handling) is designed as a standardized template with precise functionality for its specific task. These templates can be copied and deployed across multiple applications without modification, achieving both manufacturing precision for the specific task and adaptability to different applications through configuration rather than redesign.
Data Source
AI summary
A method for conducting a conversation between a user and a virtual agent is disclosed. The method includes receiving, by an ASR sub-system, a plurality of utterances from the user, and converting, by the ASR sub-system, each utterance of the plurality of utterances into a text message. The method further includes determining, by a NLU sub-system, an intent, at least one entity associated to the intent, or a combination thereof from the text message. The method further includes generating, by a conversation management sub-system, a predefined conversation map for a determination of at least one text based agent response for the text message based on: selecting at least one conversation block sub-system from a plurality of conversation block sub-systems; sending a request along with the intent and the at least one entity to each selected at least one conversation block for execution of a specific function associated with the selected at least one conversation block; receiving an agent intent from the selected at least one conversation block sub-system and generating the at least one a text based agent response for the user based on the at least one agent intent.


