Real-time AI agent for operational decision-making
AI-powered real-time co-pilot for operations teams.
It detects anomalies, explains root causes, simulates actions, and learns from outcomes — all in a beautiful dashboard.
| Feature | Description |
|---|---|
| ⚡ Event Ingestion | Simulates real-time order, revenue, and inventory data |
| 🔍 Anomaly Detection | Detects spikes/drops using Z-score logic |
| 🧠 LLM Reasoning | (Mocked) natural explanations of anomalies |
| 🛠️ Decision Engine | Suggests actions (reroute, boost ads, flag for audit) |
| 📤 Action Layer | Sends Slack alerts, simulates API postbacks |
| 🧾 Memory | Logs anomalies + actions in CSV |
| 🌀 Feedback Loop | Tracks if actions were effective (pending, success, failed) |
| 📊 Analytics | Charts performance in Streamlit UI |
| 🔁 Autonomous Loop | Runs every 10 seconds in the background |
git clone https://github.com/your-username/liveops-agent.git
cd liveops-agent
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python data/simulate_stream.py # In one terminal
python auto_agent.py # In another
streamlit run ui/dashboard.py # In browser