Context-Aware AI Querying for Enterprise Financial Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing applications require users to navigate complex menu hierarchies and tool selections to perform tasks with datasets, leading to inefficiencies and limited functionality, as shortcuts are constrained by screen space and users often lack awareness of full menu hierarchies.
Innovation Solution
A computer-implemented method using a large language model (LLM) generates context-specific prompts based on user roles and domain-specific knowledge, retrieving data from databases to provide efficient and accurate responses to natural language queries, leveraging Retrieval Augmented Generation (RAG) to enhance query understanding and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If applications provide deep, rich functionality through menu hierarchies and tool selections, then the functionality and versatility are improved, but the ease of operation and efficiency deteriorate due to complex navigation requirements
Solution Approach 1:
The patent introduces a natural language processing intermediary that mediates between the user and the complex application interface. Users can query data using simple natural language instead of navigating through multiple menu hierarchies, while the system translates these queries into the appropriate data retrieval operations, thus preserving full functionality while dramatically improving ease of operation
Solution Approach 2:
The patent replaces the mechanical interaction model of clicking through menu hierarchies and making item selections with a natural language processing system. This substitution allows users to interact with the application's full functionality through conversational queries, eliminating the need to understand or navigate complex interface structures
2Ease of operation
If applications provide shortcuts for frequently used actions, then the ease of operation and efficiency are improved, but the functionality is limited by screen space constraints
Solution Approach 1:
The patent implements a universal natural language interface that can handle any data query functionality, replacing the need for multiple specialized shortcuts. This single interface provides multi-functional capability to access all data types and operations that would otherwise require numerous separate shortcuts, thus improving both efficiency and functionality simultaneously
Solution Approach 2:
The patent moves the interaction paradigm from a two-dimensional screen interface with limited shortcut space to a natural language dimension where users can express any query regardless of screen real estate. This dimensional shift allows unlimited functionality to be accessed through conversational interfaces without being constrained by display area
3Adaptability or versatility
If users are provided with comprehensive menu hierarchies and tools, then the functionality is improved, but the time required to complete tasks increases due to navigation complexity
Solution Approach 1:
The patent implements preliminary action by pre-processing and understanding user intent through natural language before any navigation or data retrieval occurs. The system analyzes the query, determines the required data and operations, and executes them directly, eliminating the time-consuming navigation through menu hierarchies that would otherwise be required to access the same functionality
Data Source
AI summary
Systems, methods, and computer-readable media provide a context-specific prompt to answer a user query. The systems, methods, and computer-readable media determine a context based on content of a natural language request and/or determine a role of a user who submitted the natural language request. Additionally or alternatively, templates or RAG sources that will be used for prompt generation may include financials domain-specific knowledge or other domain-specific knowledge or insights. Inclusion of this additional information in the prompt enhances the context to promote more accurate results from a large language model. In one embodiment, the prompt templates are created from various RAG sources, such as payables, general ledger, receivables, and asset management, containing structured data and information specific to the financial domain, enterprise, or other domain, which helps craft accurate prompts. A prompt is generated that identifies a subset of available fields and other selected information based on the role or other context. The prompt template may contain domain-specific knowledge uses a relevant domain or enterprise information to drive relevant results, and an executable query is generated by a large language model based on the prompt. The executable query causes data to be retrieved from a database to generate a result, and information is displayed based at least in part on the result.