December 2025 – 3
We are pleased to announce the release of the OracleChat Agentic System Version 1.0, a major AI-driven enhancement to the BluB0X platform that significantly improves how users interact with building and platform data.
Release Overview:
Release Version: 1.0
Release Date: December 2025
Target Users: BluB0X Platform Role Users
Platform: Azure Functions (Python 3.11)
Executive Summary:
The OracleChat Agentic System introduces an intelligent, multi-agent architecture that autonomously reasons over user queries and dynamically selects the appropriate tools. It seamlessly combines SQL-based database queries with document-based RAG (Retrieval-Augmented Generation) to deliver accurate, contextual, and real-time responses.
Key Highlights:
- Intelligent Agent System: Autonomous reasoning to select the right tools per query
- Multi-Source Intelligence: Unified SQL + RAG-based answers
- Real-Time Streaming: Live progress and response streaming via SignalR
- Specialized Agents: Building-level Agent & BluSKY Platform-level Agent
- Advanced Tools:
- Building SQL Tool
- Device SQL Tool
- Building RAG Tool
- BluSKY Platform RAG Tool
- Enhanced Security: Read-only DB access, query validation, safety controls
- Production-Ready: Dockerized deployment with comprehensive testing
PBIs Delivered (Summary):
All planned PBIs for this release have been successfully completed, including:
- Basic Agentic System Implementation – Autonomous tool selection for complex queries
- Real-Time Frontend Integration – Progressive updates and live status messages
- Device-Specific Intelligence – Deep insights into access control, cameras, elevators, intercoms, and more
- Platform-Level Intelligence – Queries across BluSKY features, tickets, and documentation
- RAG Pipeline v1 & Deep RAG – Accurate, cited, and context-aware documentation answers
- Content Classification Enhancements – Improved handling of documents with images
- Critical Fixes & Validation – Robust error handling and validation
- RAG Status Tracking & Intelligent Chunking – Better indexing visibility and relevance
- Dead Letter Queue Handling – Reliable capture and reprocessing of failed requests
What Can You Ask?
Building Agent Examples
- “How many occupants are in this building?”
- “List all cameras on Floor 2.”
- “Find persons in the IT department and their access rights.”
BluSKY Agent Examples:
- “How does BluSKY handle visitor management?”
- “What are the latest BluSKY platform features?”
- “Explain BluSKY’s access control capabilities.”
Security & Compliance:
- Read-only database access
- SQL injection prevention and query sanitization
- Building-scoped data access
- Enterprise-grade security and compliance controls
Documentation & Support:
A comprehensive documentation suite is available, covering system architecture, agent behavior, APIs, security configuration, and onboarding guides.