External Knowledge Assets for Context-Aware AI Chatbots
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI/ML chatbots are limited by their stateless nature and lack of customization for specific organizational contexts, leading to inaccurate and contextually irrelevant responses due to insufficient training on internal terminology and data.
Innovation Solution
A computer system enables organizations to customize chatbots by providing external knowledge assets, such as knowledge bases, that include definitions, organizational hierarchies, and glossaries, allowing AI/ML models to interpret user prompts more accurately and generate contextually relevant responses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If generalized AI/ML models are used, then the chatbot can handle a wide range of topics, but it cannot be customized for particular organizational uses or contexts
Solution Approach 1:
The system segments the chatbot configuration into modular components: core AI/ML model, knowledge base, and organizational context data. This allows the base model to remain generalized while organizational-specific customization is achieved through separate, independently configurable knowledge bases that can be added or modified without changing the core model.
Solution Approach 2:
The system performs preliminary action by pre-configuring knowledge bases with organizational terminology, glossaries, and context data before the chatbot begins operation. This allows the chatbot to be customized for particular organizational uses in advance, eliminating the need for complex post-deployment customization while maintaining versatility.
2Measurement precision
If the chatbot uses internal organizational terminology and jargon, then it can provide more contextually relevant responses, but it requires additional training data and computing resources
Solution Approach 1:
The system introduces a knowledge base as an intermediary between the generalized AI/ML model and organizational terminology. The knowledge base stores glossaries, definitions, and context data that enable the chatbot to understand internal jargon without requiring extensive training of the core model, thus improving response accuracy while avoiding the computational cost of retraining large models.
Solution Approach 2:
The system creates a copy of organizational knowledge and terminology in the form of structured knowledge bases that can be independently updated and maintained. This allows the chatbot to access organizational context through lightweight knowledge representations rather than requiring the expensive process of training the AI/ML model itself on organizational data.
3Reliability
If the chatbot is updated with new knowledge assets, then it can improve recall and accuracy, but it requires time and resources to retrain the model
Solution Approach 1:
The system segments the knowledge updating process from the model training process. Knowledge assets such as glossaries, definitions, and organizational data are stored in separate knowledge bases that can be independently updated without triggering model retraining. This allows rapid updates to improve chatbot effectiveness while avoiding the time-consuming model retraining process.
Solution Approach 2:
The system enables self-service updating of knowledge bases, where administrators can add, modify, or remove knowledge assets without requiring technical expertise in model training or deployment. This allows organizations to continuously improve chatbot recall and accuracy by updating knowledge bases independently, eliminating the time loss associated with model retraining cycles.
Data Source
AI summary
Methods, systems, and apparatus, including computer-readable media, for artificial intelligence chatbots using external knowledge assets. In some implementations, a system stores a knowledge base that comprises one or more knowledge items for an organization. The system receives a user prompt for a chatbot, and generates a chatbot response to the user prompt using one or more artificial intelligence and/or machine learning (AI/ML) chatbots. The chatbot response to the user prompt is generated at least in part based on the one or more AI/ML models processing the one or more knowledge items from the knowledge base. The system provides the chatbot response for presentation.