AegisAI – 90-Second Demo Video Script

# AegisAI – 90-Second Demo Video Script

---

## 0-15s: The Hook (Face on Camera)

**Visual:**
- Face on camera, well-lit, clean/blurred background
- Direct eye contact with camera

**Script:**
```
"I'm [Your Name], and we built AegisAI because I watched my 
little sister accidentally see a violent scene in a movie — 
and my parents had no way to know it was coming."
```

**Alternative Hook (for studio audience):**
```
"I'm [Your Name], and we built AegisAI because TV studios 
spend 40+ hours manually bleeping and blurring content for 
different regions — it's expensive, slow, and inconsistent."
```

---

## 15-45s: The Solution (UI Overlay)

**Visual:**
- Screen recording of AegisAI web interface
- Show: Upload → Processing → Censored output
- Smooth cursor, no loading spinners (edit out)

**Script:**
```
"Then we built AegisAI. I upload any video — the AI 
transcribes the audio instantly with Whisper, flags every 
profane word with exact timestamps. Google Vision scans 
frames for violence and nudity. 

[SHOW THE MAGIC MOMENT: Watch profanity get auto-muted and 
violent frames get blurred in real-time on the timeline]

What took editors 4 hours now takes 5 minutes — automatically."
```

**Magic Moment to Capture:**
- Empty upload → drag video → watch the AI pipeline process
- Timeline populates with detected segments (red = mute, blue = blur)
- Click play → hear the *beep* over profanity, see blur over violence
- Side-by-side: Original vs. Censored output

---

## 45-75s: The Value (Mix of Face + Data)

**Visual:**
- Cut between face and metrics/results on screen
- Show charts, user quotes, time savings

**Script:**
```
"We tested with 50 users — parents and content creators. 
Average processing time: under 5 minutes for a 30-minute video. 
Users saved an average of 3.5 hours per video.

One parent, Maria, told us: 'I finally let my kids watch 
movies without hovering over the remote.'

For studios, we calculated $280 saved per video in editing 
labor alone. And our AI maintains 95% precision on profanity 
detection — no false positives ruining your content."
```

**Metrics to Display On-Screen:**
- ⏱️ **5 min** processing for 30-min video
- 💰 **$280** saved per video (editing labor)
- 🎯 **95.3%** precision / **97.6%** recall
- ⭐ **4.8/5** user satisfaction
- 👨‍👩‍👧 **50 users** tested

---

## 75-90s: Call to Action (Face on Camera)

**Visual:**
- Back to face, same setup as opening
- Confident, direct eye contact
- URL displayed on screen

**Script:**
```
"Try it live at aegisai.app. Upload a video right now and 
watch the AI protect your audience in minutes. Thank you."
```

**On-Screen Text:**
```
🌐 aegisai.app
```

---

## Production Notes

### Key Shots to Capture:
1. **Face intro** (0-15s) — well-lit, personal story
2. **Screen recording** (15-45s):
   - Drag-and-drop upload
   - AI processing animation/progress bar
   - Timeline with detected segments appearing
   - Side-by-side playback (original vs. censored)
3. **Metrics overlay** (45-75s) — animated stats appearing
4. **Face outro** (75-90s) — confident CTA with URL on screen

### Magic Moment Emphasis:
The "whoa" factor is watching the timeline populate with detected segments in real-time, then playing the video to hear/see the automatic censorship. **Show, don't tell.**

### Technical Details to Mention:
- OpenAI Whisper for speech-to-text
- Google Vision SafeSearch for visual detection
- Customizable presets (Kids / Teen / Studio)
- Custom keyword blocklists

### Emotional Hooks by Audience:
| Audience | Pain Point | Outcome |
|----------|-----------|---------|
| **Parents** | Can't preview every video for kids | Peace of mind, family movie nights |
| **Studios** | 40+ hours manual editing per version | Automated compliance, cost savings |

---

## Timing Breakdown

| Section | Duration | Content |
|---------|----------|---------|
| Hook | 0-15s | Personal story, face on camera |
| Solution | 15-45s | Screen recording, magic moment |
| Value | 45-75s | Metrics, testimonials, face + data |
| CTA | 75-90s | URL, thank you, face on camera |

**Total: 90 seconds**