docs/DEMO_SCRIPT.md

🎥 GUVI Demo Video Script (1 Minute)

Goal: Convince judges that this is a working, improved, and enterprise-ready honeypot system.

---

🎬 Scene 1: Introduction (0:00 - 0:10)

• Visual: Show the README.md with the "Sentinel Scam Honeypot" banner and badges.
• Voiceover: "Hi, this is Team [Team Name]. We've built Sentinel, an autonomous AI honeypot that traps scammers and extracts critical intelligence for Indian law enforcement."

🎬 Scene 2: The Attack (0:10 - 0:25)

• Visual: Split screen. Left: Your API/Dashboard. Right: Swagger UI (/docs).
• Action: Send a simulated scam message via Swagger UI (/api/v1/analyze).
  • Input: "Congratulations! You won 1 crore lottery! Send tax amount to verify at winner@paytm or call +91-9876543210"
• Voiceover: "Here, we simulate a typical lottery scam message attacking our citizen hotline."

🎬 Scene 3: The AI Response (0:25 - 0:40)

• Visual: Show the JSON response in Swagger. Highlight:
  • scam_detected: true
  • honeypot_response: (e.g., "Wah! Sach mein?")
  • extracted_intelligence: Phone number and UPI ID.
• Voiceover: "Our hybrid AI instantly detects the scam, identifying it as a 'Lottery Fraud'. It generates a believable response using our 'Sharma Uncle' persona to hook the scammer, while automatically extracting their UPI ID and phone number."

🎬 Scene 4: Enterprise Intelligence (0:40 - 0:50)

• Visual: Switch to the Streamlit Dashboard (dashboard.py). Show the graphs.
• Action: Point to "Top Scams" or "Live Threat Feed".
• Voiceover: "All data is aggregated in real-time. We cluster campaigns and generate threat intelligence reports compatible with CERT-In and NPCI formats."

🎬 Scene 5: Conclusion (0:50 - 1:00)

• Visual: Show the ARCHITECTURE.md diagram or a slide with "1.4 Billion Scale".
• Voiceover: "Sentinel is built to scale for 1.4 billion citizens using an asynchronous agentic architecture. We don't just block calls; we waste scammer time and stop the money flow."

---

🛠️ Validation Checklist for Video

1. Are badges visible? (Shows polish)
2. Is the JSON output readable? (Proofs it works)
3. Is the Dashboard active? (Visual appeal)
4. Did you mention "Citizen Safety"? (Emotional hook)