Connected AI Agents for Permissioned Vector Search in Messaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI models, such as Large Language Models (LLMs), are limited in responding to questions involving personal and enterprise content items due to their training data, and users face difficulties in locating relevant information within messaging applications, leading to inefficient and distracting search processes, especially on mobile devices.
Innovation Solution
A connected AI agent system that integrates a connector service and AI agent across multiple servers, vectorizes user content items, and executes agent objects to provide personalized responses within messaging applications, leveraging user permissions and management policies to access and respond to queries using vector databases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users perform text searching in folders or message histories to locate relevant personal and enterprise items, then they can find information, but the process is time-consuming and requires leaving the messaging environment
Solution Approach 1:
The system pre-processes and vectorizes content items from personal and enterprise sources before they are needed for searching. By converting documents, emails, and other content into vector representations in advance and storing them in vector databases, the system eliminates the need for time-consuming text searching when users need information, as AI agents can directly query the pre-processed vectors.
Solution Approach 2:
The patent replaces traditional text-based mechanical search methods with AI-driven semantic search using vector databases. Instead of manually searching through text in folders or message histories, the system uses AI agents to perform semantic queries on vectorized content, significantly reducing search time while improving accuracy through understanding of meaning rather than keyword matching.
2Adaptability or versatility
If users open multiple different applications to perform searches and access content, then they can find information, but their attention is distracted and resources are consumed
Solution Approach 1:
The system merges multiple content sources (personal documents, enterprise documents, emails, text conversations) and multiple functions (searching, AI analysis, response generation) into a unified messaging environment. AI agents operate within the same interface where users communicate, eliminating the need to switch between applications while maintaining access to diverse content types and advanced search capabilities.
Solution Approach 2:
The messaging application with integrated AI agents becomes a universal platform that performs multiple functions: traditional messaging, semantic search across personal and enterprise content, AI-powered information retrieval, and automated response generation. This multi-functional system replaces the need for separate specialized applications, improving ease of operation while maintaining versatility.
3Adaptability or versatility
If public LLMs are used to answer questions, then they provide general knowledge, but they cannot access personal and enterprise content items
Solution Approach 1:
The system introduces AI agents as intermediaries between public LLMs and private enterprise content. The AI agents retrieve relevant information from vector databases containing personal and enterprise content, then provide this information to the LLM for processing. This intermediary approach allows the LLM to answer questions about private content while maintaining the reliability benefits of using established public models for general knowledge.
Solution Approach 2:
The system segments the knowledge source into two parts: public knowledge from trained LLMs and private enterprise content from vector databases. AI agents selectively query the appropriate source based on the question type, using the LLM for general knowledge and the vector database for organization-specific information. This segmentation allows the system to maintain both the versatility of public LLMs and the reliability of access to private content.
Data Source
AI summary
Systems and methods are described for a connected AI agent for managed multidimensional search based on an electronic message and management policies. A messaging application at a client device can send a new electronic message, such as an email, to a connector service. An attachment can be ingested and stored in a vector database. Then one or more artificial intelligence (“AI”) agents can be selected for responding to the body of the email, such as a query in the body. The responses can be formatted and sent to multiple parties, such as a sending user of the electronic message and a recipient that was copied or also sent the new electronic message. The AI agents can use different AI models, prompts, and vector databases depending on user permissions. This allows for building up vector databases with relevant content items and answering user questions based on those vector databases.