Industry: Medical Technology / Healthcare Infrastructure
Built for a medical tech company to help non-technical clinic managers oversee critical battery systems without needing to learn complex database tools.
Problem
Existing dashboards were too technical for the average user, making it hard to find quick answers.
Complex Dashboards: Users had to navigate deep menus and filters to understand battery health.
High Support Load: Clients often called support teams to ask simple questions about their data.
Slow Decision-Making: Identifying "at-risk" batteries required manual data analysis and calculation.
Solution
A smart chatbot (AI Agent) that talks directly to the database and answers questions in plain English.
Natural Language Interaction: Users ask questions like "Which batteries will fail this week?" and get instant answers.
ReAct Agent Architecture: A system that "thinks" and selects the right tools to calculate trends or check thresholds.
Custom Tooling: 14 specialized tools that bridge the gap between the AI and the existing SQL database.
Direct Integration: The agent lives inside the current dashboard as a simple chat module via FastAPI.
Tech Stack
AI Frameworks:Azure OpenAI and LangGraph for agent logic.
Architecture:ReAct Agent pattern with automated tool calling.
Backend:Python and FastAPI for high-performance API endpoints.
Data: Integration with SQL Stored Procedures for live data access.
Results & Impact
Self-Service Data: Clinic staff can now get reports without help from technical teams or support staff.
Proactive Maintenance: Faster identification of batteries below 80% capacity or those at risk of failure.
Better Visibility: Improved oversight of thousands of assets across hundreds of different locations.
User Adoption: Increased platform value by making complex data accessible to everyone.