Virtual Assistant Server Intent Routing via Context Prioritization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing approach to chatbot development leads to user dissatisfaction due to the complexity of interacting with multiple chatbots, each with different capabilities and conversation styles, and results in resource wastage for enterprises due to low standardization and inefficient routing of conversations in universal bots, caused by similarities in training data sets and varying training sizes among child bots.
Innovation Solution
A method and device for orchestrating automated conversations using a virtual assistant server that evaluates utterances to identify intents, calculates common scores, ranks them, and prioritizes based on context information to select the appropriate child bot for response, ensuring accurate routing and enhancing user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple specialized chatbots are developed by different departments, then each chatbot can be optimized for its specific domain, but users must interact with multiple chatbots with different capabilities and conversation styles, increasing complexity and reducing user satisfaction
Solution Approach 1:
The patent implements a universal bot that can perform multiple functions by routing user queries to different specialized child bots based on intent recognition. The universal bot serves as a single entry point that handles various domain-specific tasks, eliminating the need for users to interact with multiple separate chatbots while maintaining domain specialization through the child bot architecture.
Solution Approach 2:
The universal bot acts as an intermediary between users and multiple specialized child bots. It receives user queries, evaluates intents using natural language processing, and routes appropriate queries to the relevant child bots. This intermediary layer abstracts the complexity of multiple specialized systems into a single unified interface for users.
2Adaptability or versatility
If child bots are trained with similar training utterances, then they can all respond to common queries, but routing decisions become ambiguous when multiple child bots send responses to the same utterance
Solution Approach 1:
The system implements a feedback mechanism where child bots send their responses and confidence scores back to the universal bot. The universal bot uses this feedback information, particularly the confidence scores, to make informed routing decisions. When multiple child bots respond to the same utterance, the universal bot evaluates the confidence scores and selects the most appropriate child bot, resolving routing ambiguity.
Solution Approach 2:
The patent changes the parameter used for routing decisions from simple match criteria to confidence scores. Each child bot provides a confidence score indicating its level of certainty about its response, and the universal bot uses this quantitative parameter to make objective routing decisions. This parameter change transforms subjective routing ambiguity into an objective scoring process.
3Adaptability or versatility
If child bots have varying training data set sizes, then they can be optimized for different complexity levels, but routing decisions may be unreliable when child bots have small training data sets
Solution Approach 1:
The confidence score feedback mechanism allows the universal bot to assess the reliability of each child bot's response based on its training data characteristics. Child bots with larger, more comprehensive training data sets typically provide higher confidence scores, enabling the universal bot to reliably distinguish between well-trained and less-trained child bots and make accurate routing decisions accordingly.
4Adaptability or versatility
If enterprises develop multiple specialized chatbots, then they can address specific departmental needs, but it causes waste of time and resources due to multiple development and deployment cycles and low standardization
Solution Approach 1:
The universal bot architecture provides a standardized platform that can be used across all departmental chatbots. Instead of developing completely separate chatbot systems for each department, enterprises can build upon the universal bot framework, sharing common components such as intent recognition, routing mechanisms, and user interface management. This significantly reduces development time and resources while maintaining departmental specialization.
Solution Approach 2:
The system segments the chatbot functionality into a universal bot framework and separate child bots for different domains. This segmentation allows for standardized development of the universal framework while enabling customized child bots for specific departments. The modular architecture facilitates reusable components and reduces the need to reinvent the wheel for each departmental chatbot.
Data Source
AI summary
A virtual assistant server receives an utterance from an input mechanism. Upon receiving the utterance, the virtual assistant server, evaluates the utterance to identify a plurality of intents corresponding to the utterance and calculates common scores using natural language processing techniques for each of the identified plurality of intents. The virtual assistant server ranks the identified plurality of intents based on the calculated common scores and based on the ranking identifies a first winning intent and a second winning intent. Subsequently, the virtual assistant server prioritizes one of the first winning intent or the second winning intent to identify a final winning intent based on context information. The virtual assistant server executes the final winning intent and forwards a response to the input mechanism.