Multi-Level NLP Framework for Query Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing customer support systems face inefficiencies in handling natural language queries, leading to long wait times and unsatisfied customers due to the complexity of question-answer databases and incorrect routing of queries.
Innovation Solution
A computer-implemented method using a multi-level framework of natural language models to determine the relevance of question-answer domains and provide appropriate responses, including the use of top-level and lower-level models to analyze queries and identify matching metrics for efficient answer provisioning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a large and complex question-answer database is used to cover all possible customer questions, then the system can provide comprehensive answers, but the complexity of the database increases and query processing becomes slower
Solution Approach 1:
The patent segments the large question-answer database into multiple topic-specific databases (e.g., billing, account management, product information). Each database is managed independently, reducing the complexity of searching through the entire database while maintaining comprehensive coverage through selective querying of relevant topic databases.
Solution Approach 2:
The patent introduces a query routing component that acts as an intermediary between the customer query and the question-answer databases. This routing component analyzes the query, determines the relevant topic, and directs the query to the appropriate specialized database, reducing overall system complexity while maintaining comprehensive answer capabilities.
2Reliability
If the enterprise searches through the entire question-answer database for every customer query, then all possible answers can be found, but the query processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing and categorizing questions into topic-specific databases before customer queries arrive. Query routing rules are pre-established for different question types, enabling rapid identification of relevant databases without searching the entire question-answer repository for each customer query.
Solution Approach 2:
The patent divides the monolithic question-answer database into segmented topic-specific databases (billing, accounts, products, etc.). This segmentation allows the system to search only the relevant subset of databases based on query type, significantly reducing processing time while maintaining answer accuracy through targeted searching.
3Ease of operation
If customer queries are transferred to the correct section when the initial section is wrong, then the customer receives appropriate assistance, but the customer experiences long waiting periods
Solution Approach 1:
The patent introduces an automated query routing system as an intermediary that immediately directs customer queries to the correct department or agent based on pre-established routing rules. This eliminates manual transfer delays and waiting periods while ensuring customers reach the appropriate service section.
Solution Approach 2:
The patent performs preliminary routing analysis and department identification before customer queries are transferred. By pre-configuring routing rules for different query types and analyzing queries in advance, the system prepares the correct destination ahead of time, eliminating wait periods associated with manual transfer and department identification.
4Measurement precision
If a multi-level framework of natural language models is used to analyze queries, then query understanding accuracy improves, but the computational complexity and processing time increase
Solution Approach 1:
The patent segments the multi-level natural language model framework into distinct analysis layers (syntax analysis, semantics analysis, intent recognition, entity extraction). Each layer processes specific aspects of the query independently, improving understanding accuracy through systematic analysis while managing complexity through modular organization of model components.
Solution Approach 2:
The patent applies partial analysis actions by selectively engaging different levels of the natural language model framework based on query complexity. For simple queries, only essential analysis layers are activated, reducing computational complexity while maintaining sufficient understanding accuracy. For complex queries, the full multi-level framework is engaged.
Data Source
AI summary
A computer-implemented method and an apparatus manage natural language queries of customers. A natural language query provided by a customer on an enterprise interaction channel is received. The natural language query is analyzed to determine if an answer to the natural language query exists in at least one question-answer (QA) domain from among a plurality of QA domains by analyzing each QA domain from among the plurality of QA domains using a multi-level framework of natural language models. An answer to the natural language query is provided to the customer on the enterprise interaction channel if such an answer in available in the plurality of QA domains. If an answer is not available, then an appropriate response is provided to the customer to assist the customer.