AI Assistant Configuration Using Reusable Procedural Knowledge
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI-enhanced RPA systems require high-level programming skills and are inefficient for configuring multiple language-based tasks, necessitating manual rewriting for each new service or procedure, limiting widespread adoption.
Innovation Solution
A method for configuring an AI assistant using a generic master program and separate procedural knowledge information, allowing service builders to describe tasks in a structured descriptive language without programming skills, enabling flexible and efficient service configuration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If proprietary program instructions are used to configure AI-enhanced RPA systems, then the system can orchestrate workflows with AI GPT agents, but high-level programming skills are required and the configuration process becomes complex and inefficient
Solution Approach 1:
The patent introduces an intermediary layer (the configuration system with natural language processing) between the user and the proprietary program instructions. This intermediary translates user-friendly natural language descriptions into the required program instructions, eliminating the need for users to directly write or understand complex code while still enabling workflow orchestration capabilities
Solution Approach 2:
The patent replaces the mechanical system of manual programming and code writing with an automated natural language processing system. Instead of requiring users to mechanically write program instructions, the system automatically generates them from natural language inputs, substituting the manual coding process with an intelligent translation process
2Adaptability or versatility
If manual setup of AI instructions is performed for each new service, then the service can be customized, but the instructions require rewriting for each new service or procedure which is inefficient and not scalable
Solution Approach 1:
The patent implements preliminary action by creating a reusable template structure that pre-defines the framework for service configuration. This template includes predefined elements and relationships that can be automatically instantiated for new services, eliminating the need to manually rewrite entire instruction sets while still allowing customization through parameter specification
Solution Approach 2:
The patent uses copying by allowing service configurations to be replicated from templates or existing service definitions. Once a service is configured once, the configuration can be copied and adapted for similar services, dramatically reducing the time and effort required for repeated configuration tasks while maintaining service-specific customization
3Extent of automation
If autonomous AI agents execute provided software code, then the agent can perform tasks, but programming skills from an expert are required to write the code
Solution Approach 1:
The patent enables self-service by allowing the system to automatically generate the required program instructions from natural language descriptions provided by non-expert users. The system serves itself by translating high-level user requirements into executable code without requiring external programming expertise, making task execution capability accessible to a broader audience
Data Source
AI summary
A computer-implemented method for providing configuration information to an Artificial Intelligence, AI, assistant to cause the AI assistant to provide a response to a user request received through a user interface comprises the steps of obtaining first instructions to get the AI assistant to retrieve a master program comprising pseudo-code instructions to be charged into the AI assistant, the pseudo-code instructions being configured to cause the AI assistant to run the sequence of tasks defined in procedural knowledge information, external to the master program, obtaining second instructions to get the AI assistant to retrieve said knowledge procedural information separately from said master program, said knowledge procedural information being specific to the service and comprising a structured description of a sequence of tasks defining said service, and providing said first and second instructions for transmission to said AI assistant through a communication interface.