Conversation-Based Search System for Enterprise Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional insight search systems are primitive, requiring users to know the context and specific models for searching, lacking user understanding, and producing minimal, unreadable outputs.
Innovation Solution
A conversation-based search system that processes verbal user input to determine contexts, provides data scope mapping, executes queries, and displays insights as visualizations within a story, using natural language libraries and data fencing to access relevant enterprise data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional text-based search with prebuilt modeled objects is used, then search functionality is provided, but user accessibility and ease of operation deteriorate because users need to know context and specific models
Solution Approach 1:
The patent introduces a natural language processing intermediary layer that translates casual user queries into structured search queries. This mediator handles the complexity of context understanding, model selection, and parameter mapping automatically, allowing users to interact without knowing the underlying system complexity.
Solution Approach 2:
The system performs self-service by automatically determining the context, selecting appropriate models, and identifying relevant parameters based on the user's natural language query. This eliminates the need for users to manually specify these technical details, significantly improving ease of operation.
2Loss of information
If traditional search systems operate with expectation that user is fully aware of what they are looking for, then query processing is simplified, but user understanding and insight quality worsen
Solution Approach 1:
The system performs preliminary actions by proactively analyzing the user's intent, determining the appropriate context before the search is executed. This includes identifying the business domain, selecting relevant models, and pre-processing the query to extract key parameters, thereby reducing information loss.
Solution Approach 2:
The system implements feedback mechanisms to verify and refine its understanding of user intent. By analyzing the query results and user interactions, the system can adjust its context determination and provide more accurate insights, reducing information loss while managing processing complexity.
3Productivity
If traditional search systems provide minimal output, then processing speed is improved, but output quality and readability deteriorate
Solution Approach 1:
The patent segments the output generation process into multiple stages: query processing, context determination, model selection, parameter extraction, and result formatting. This segmentation allows the system to maintain processing speed by parallelizing independent tasks while ensuring comprehensive insight quality through systematic processing at each stage.
4Ease of operation
If verbal input processing and context determination are added, then user accessibility is improved, but system complexity increases
Solution Approach 1:
The patent implements a universal natural language processing module that handles multiple functions: speech-to-text conversion, query interpretation, context determination, and parameter extraction. This multi-functional approach improves user accessibility while managing system complexity by consolidating capabilities into a single versatile component rather than separate specialized systems.
Data Source
AI summary
Methods, systems, and computer-readable storage media for receiving, by a conversation-based search system (CSS) of an analytics system, verbal input from a user, providing, by the CSS, text input based on the verbal input, processing, by the CSS, the text input to determine a set of contexts, each context in the set of context representing one or more operations of an enterprise, determining, by the CSS, one or more insights based on the set of contexts, each insight representative of a performance of the enterprise, and displaying, by the analytics system, a story comprising one or more visualizations, each visualization depicting at least one insight.