AI Copilot Task Planning for Internal and Legacy Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional copilot services struggle to efficiently utilize internal information and legacy systems within specific organizations like enterprises, making it difficult to adapt to diverse environments and provide effective task support.
Innovation Solution
An AI-based copilot service that calculates a task plan comprising unit tasks, interworks with sub-systems such as databases and legacy systems, and uses AI models to generate responses, allowing efficient resource utilization and flexible adaptation to organizational environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional copilot services use pre-learned general knowledge, then the service can be implemented with simple architecture, but the service cannot utilize internal information and legacy systems of specific organizations
Solution Approach 1:
The system is segmented into multiple independent components: an AI model for generating responses, a separate module for accessing internal information, and an interface for connecting to legacy systems. This segmentation allows the system to adapt to different organizational environments by configuring specific components without redesigning the entire architecture.
Solution Approach 2:
An intermediary layer is introduced between the AI model and the organizational resources (internal information systems and legacy systems). This intermediary handles the complexity of interfacing with diverse organizational systems, allowing the AI model to remain simple while still accessing specialized resources through standardized interfaces.
2Productivity
If copilot service integrates with multiple sub systems and legacy systems, then the service can efficiently utilize internal information, but the system complexity and difficulty of implementation increase
Solution Approach 1:
The system employs universal interfaces and standardized protocols for integrating with different sub systems and legacy systems. The AI model can perform multiple functions (generating responses, querying internal information, interacting with legacy systems) through a unified architecture, reducing implementation difficulty despite the diversity of integrated systems.
3Reliability
If copilot service accesses internal information and legacy systems, then the service can provide organization-specific task support, but the difficulty of detecting and measuring system configuration increases
Solution Approach 1:
The system incorporates feedback mechanisms that automatically detect and measure the configuration of internal information systems and legacy systems during runtime. The AI model receives feedback about system availability and status, allowing it to adapt its task execution strategy based on the detected environment without requiring complex pre-configuration analysis.
Data Source
AI summary
A processor-implemented method including calculating a task plan, the task plan including one or more unit tasks, based on a user request, performing the task plan in association with one or more sub systems corresponding to the one or more unit tasks, and providing an answer to the user request, based on a result of performing the task plan.