# One-Shot Prompt

**Topic**: AI in Healthcare
**Theme**: Corporate Dark
**Generated**: 2026-04-17
**Model**: MiniMax-M2.7 (one-shot generation)

---

## Prompt

Write a complete Node.js script using `pptxgenjs` (ES module, `.mjs`) that generates a professional 15-slide PowerPoint presentation about **AI in Healthcare**, themed with the **Corporate Dark** visual style.

The deck must feel like it was designed by a senior presentation designer — not a tutorial exercise. Every slide should have visual intention, data where appropriate, and a clear narrative arc.

### Theme: Corporate Dark

**Palette**:
- Primary: `1B2A4A` (deep navy)
- Secondary: `2E4A7A` (medium navy)
- Accent: `E8913A` (warm orange)
- Text: `FFFFFF` (white on dark), `2D3436` (dark on light)
- Light text: `B0BEC5` (muted silver)
- Background: `0D1B2A` (near-black navy)
- Light BG: `F5F7FA` (cool white for content slides)
- Dark: `091520` (deepest navy)

**Background strategy**:
- Title slide: solid `0D1B2A` with a thin accent bar at the bottom
- Content slides: `F5F7FA` (light) with navy title bar at top
- Section dividers: gradient from `1B2A4A` to `0D1B2A`
- Closing: match title slide

**Chart colors**: `2E4A7A`, `E8913A`, `5BA0D9`, `7EC8A0`, `D4556B`
**Table header**: `1B2A4A` fill, white text
**Table alternating**: `EDF2F7` / `FFFFFF`

---

### Slide Structure (15 SLIDES — ALL REQUIRED)

| # | Slide Type | Content |
|---|-----------|---------|
| 1 | **Title Slide** | "AI in Healthcare: Transforming Medicine in 2026" — subtitle: "From Diagnosis to Drug Discovery, the AI Revolution", date, "Prepared by MiniMax-M2.7" |
| 2 | **Agenda / Overview** | 5 bullets: Diagnostics & Imaging, Drug Discovery, Robotic Surgery, Operational Efficiency, Ethical Considerations |
| 3 | **Context / Why This Matters** | Problem statement: 10M+ diagnostic errors annually in US. Key stat callout: "1 in 10 patients affected" |
| 4 | **Key Data Point** | Large display stat: "$150B" — projected global AI healthcare market by 2030 (CAGR 40%). Supporting text about compound growth |
| 5 | **Market/Landscape Overview** | Bar chart: North America ($48B), Europe ($35B), Asia-Pacific ($42B), Latin America ($12B), Rest of World ($13B) — 2026 estimates |
| 6 | **Breakdown / Categories** | Doughnut chart: Diagnostics 30%, Drug Discovery 25%, Robotics 20%, Administrative 15%, Remote Monitoring 10% |
| 7 | **Timeline / History** | Visual timeline: 2019 (FDA first AI cleared), 2021 (AlphaFold protein folding), 2023 (Da Vinci 5 launch), 2024 (GPT-4 MedPrompt), 2025 (Autonomous imaging), 2026 (AI Agent triage) |
| 8 | **Comparison Table** | Compare 4 AI systems: IBM Watson Health, Google DeepMind, Microsoft Dragon, NVIDIA Clara — across: Accuracy, Deployment, Speciality, Cost |
| 9 | **Trend Analysis** | Line chart: AI adoption in hospitals 2020-2026: 12%, 18%, 29%, 41%, 55%, 67% |
| 10 | **Case Study / Example** | Mayo Clinic AI triage system — reduced wait times 34%, improved critical case detection 22%, stats in callout boxes |
| 11 | **Challenges & Risks** | Risk matrix: Data Privacy (High), Algorithmic Bias (High), Regulatory Gaps (Medium), Workforce Displacement (Medium), Liability (Uncertain) |
| 12 | **Opportunities / Solutions** | 4 opportunity cards: Early Diagnosis at Scale, Personalised Treatment Plans, Operational Automation, Global Health Equity |
| 13 | **Future Outlook** | Forecast 2027-2030: AI-assisted diagnosis reaches 90% accuracy, AI-designed drugs enter trials, fully autonomous surgical robots in 3 major systems |
| 14 | **Key Takeaways** | 5 numbered takeaways with icon-like circles: 1. AI augments, not replaces clinicians; 2. Data quality determines AI quality; 3. Regulation is catching up; 4. Bias mitigation is non-negotiable; 5. Interdisciplinary teams win |
| 15 | **Thank You / Q&A** | Closing: "Thank you" + placeholder contact + "Questions welcome" + subtle branding |

---

### Visual Quality Requirements

- No slide should be just a title + bullet list. Every slide needs a visual element (chart, shape, table, callout box, or layout composition)
- Use the full slide canvas — don't cluster everything in the top-left
- Consistent margins: 0.5" from edges minimum
- Visual hierarchy: title → subtitle → body → footnote, each at distinct sizes
- Title: 28-36pt, bold, theme primary color (or white on dark backgrounds)
- Subtitle: 18-22pt, regular weight, secondary color
- Body: 14-16pt, dark grey or theme text color
- Footnotes/sources: 9-10pt, light grey
- Never more than 3 font sizes on one slide
- Use `fontFace: "Arial"` as the safe default

### Shape-Based Design

- Colored rectangles as section headers, accent bars, and background blocks
- Rounded rectangles as "cards" for key points
- Circles or ellipses as "icon" placeholders with numbers or letters inside
- Lines as dividers and connectors
- Arrows to show flow or direction
- Use `rectRadius` for modern rounded corners

### Chart Quality

- Every chart needs a clear title
- Use the theme's color palette for chart colors (3-5 colors max)
- Show data labels on charts
- Legend positioned where it doesn't crowd the data

### Speaker Notes

Every slide MUST include speaker notes with:
- 2-3 talking points per slide
- Transition phrases to the next slide
- Additional context not shown on the slide

---

### Technical Constraints (NON-NEGOTIABLE)

- **Single script**: One `.mjs` file that generates the `.pptx`
- **No external images**: Use PptxGenJS shapes, charts, gradients, and text only
- **No templates**: Build every slide from scratch programmatically
- **Runnable**: `node generate.mjs` produces the `.pptx` file. Must work first try.
- **ES modules**: Use `import` syntax (`.mjs` extension)
- **NO TRIMMING**: Write every line of code. Never truncate, abbreviate, elide, or replace code with placeholder comments like `// ... rest of slides`. Every slide definition, every shape placement, every text block must be written out in full.

---

### Output

Produce `presentation.pptx` as the output file.

Run with:
```bash
npm install pptxgenjs
node generate.mjs
```