Orchestrate AI is an autonomous Support Triage Agent designed to handle complex customer queries for HackerRank, Claude, and Visa. By leveraging Retrieval-Augmented Generation (RAG), the agent provides grounded, fact-based responses while maintaining strict safety guardrails for high-risk issues.
- Autonomous Triage: Automatically classifies tickets into
Replied(grounded answers) orEscalated(security, billing, or complex issues). - RAG Architecture: Processes a knowledge base of 769 articles to ensure 100% grounded responses with automated citations.
- Dual LLM Support: Seamlessly integrated with Google Gemini 2.0 Flash and Groq (Llama 3.3) for extreme speed and reliability.
- Premium Dashboard: A sleek, Glassmorphism-styled command center with real-time triage metrics and a live activity feed.
- Batch Processing: Capable of processing hundreds of tickets in seconds via CSV automation.
- Backend: Python 3.10+, Flask, Pandas
- AI/ML: Google Generative AI (Gemini), Groq API (Llama 3.3)
- Frontend: Vanilla JavaScript (ES6+), Modern CSS (Glassmorphism), FontAwesome
- Environment: Dotenv for secure secret management
git clone https://github.com/monishad07/orchestrate-ai-support.git
cd orchestrate-ai-supportpip install -r code/requirements.txt
pip install flask flask-cors groqCreate a .env file in the root directory and add your API keys:
GOOGLE_API_KEY=your_gemini_key
GROQ_API_KEY=your_groq_keypython dashboard/server.pyOpen http://127.0.0.1:5000 in your browser.
- Retrieval: When a query enters, the system performs a semantic search over the local Markdown corpus.
- Analysis: The LLM analyzes the query against the retrieved context and applies Safety Guardrails.
- Action:
- If a solution is found: The agent generates a cited response.
- If a risk is detected (e.g., Billing/Security): The agent Escalates the ticket to a human manager.
Built with ❤️ for the HackerRank Orchestrate Hackathon.
