A multi-step workflow agent that retrieves data, reasons, takes actions, writes results and then notifies user.
Enterprise Automation Agent is a multi-step agent framework that retrieves business data, performs LLM-powered reasoning, executes actions, stores results in a database, and streams live updates to a UI via WebSockets.
This project demonstrates:
- Agent loops
- reasoning + tool calling
- workflow automation
- background workers
- event streaming
- database integration
- FastAPI backend + background worker
- Workflow:
- Retrieve data (API / DB / dummy source)
- LLM reasoning (categorize, analyze, decide)
- Execute action
- Store results
- Push updates via Websockets
- Modular agent loop
- Extensible tool system
POST /analyze
Analyzes a financial document or URL.
