# One-Shot Prompt

**Topic**: AI in Healthcare
**Theme**: Corporate Dark (deep navy, warm orange accent)
**Generated**: 2026-04-21
**Model**: kimi-k2.6:cloud

## Prompt

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

The Corporate Dark theme uses:
- Primary: `#1B2A4A` (deep navy)
- Secondary: `#2E4A7A` (medium navy)
- Accent: `#E8913A` (warm orange)
- Text: `#FFFFFF` on dark backgrounds, `#2D3436` on light backgrounds
- Light text: `#B0BEC5` (muted silver)
- Background (dark slides): `#0D1B2A` (near-black navy)
- Light background (content slides): `#F5F7FA` (cool white)
- Deep dark: `#091520`
- Chart colors: `#2E4A7A`, `#E8913A`, `#5BA0D9`, `#7EC8A0`, `#D4556B`
- Font: Arial throughout

### Narrative Arc & Slide Contents

**Slide 1 — Title Slide**
- Topic: "AI in Healthcare"
- Subtitle: "Transforming Medicine Through Intelligent Systems"
- Date, "Prepared by kimi-k2.6:cloud"
- Full-bleed dark navy background, centered white title, thin orange accent bar at bottom.

**Slide 2 — Agenda**
- 5 bullets: Market Landscape, Key Applications, Timeline, Data & Trends, Challenges & Future
- Light background, navy top bar.

**Slide 3 — Context / Why This Matters**
- Headline: Global healthcare spending projected at $18.3T by 2030, yet diagnostic error rates remain ~10%.
- Supporting text on gaps AI can fill.
- Dark background with orange stat callout.

**Slide 4 — Key Data Point**
- Giant number: "$148.4B" — Global AI healthcare market size (2023)
- Supporting text with CAGR of ~37% through 2030.
- Stat callout layout: large orange number, silver label.

**Slide 5 — Market/Landscape Overview**
- Bar chart comparing healthcare AI segments by 2023 revenue:
  - Medical Imaging: $38.2B
  - Drug Discovery: $29.5B
  - Virtual Assistants: $18.1B
  - Administrative: $24.3B
  - Robotics: $10.8B
  - Others: $27.5B
- Light background, navy chart colors.

**Slide 6 — Breakdown / Categories**
- Doughnut chart showing AI healthcare use-case distribution:
  - Diagnostics: 32%
  - Drug Discovery: 22%
  - Admin Automation: 18%
  - Virtual Care: 15%
  - Surgery/Robotics: 8%
  - Others: 5%
- Light background.

**Slide 7 — Timeline / History**
- Horizontal timeline with 5 milestones:
  - 1956: "Dartmouth — AI named as field of study"
  - 1972: "MYCIN — First expert system for antibiotic selection"
  - 1997: "Deep Blue — Proved pattern recognition at scale"
  - 2016: "AlphaFold — Protein structure prediction breakthrough"
  - 2023: "Med-PaLM 2 — LLM passes USMLE-style exams"
- Light background, navy circles connected by thin lines.

**Slide 8 — Comparison Table**
- Table comparing Traditional vs AI-Assisted approaches across 5 dimensions:
  - Diagnostic Speed, Accuracy, Cost, Scalability, Personalisation
- 4 rows: Traditional, Rule-Based Systems, ML Models, Generative AI
- Styled table with navy header, alternating light rows.

**Slide 9 — Trend Analysis**
- Line chart showing AI Healthcare Investment ($B) over years:
  - 2018: $4.1B
  - 2019: $6.3B
  - 2020: $8.2B
  - 2021: $15.1B (peak)
  - 2022: $11.4B
  - 2023: $13.8B
  - 2024: $16.2B (projected)
  - 2025: $19.5B (projected)
- Second series: "FDA AI/ML Approvals" (0, 1, 2, 5, 12, 22, 35, 48)
- Dual-axis or shared chart on light background.

**Slide 10 — Case Study / Example**
- Title: "Case Study: Mayo Clinic & PathAI"
- Three card callout boxes:
  - Challenge: High error rate in pathology review
  - Solution: AI-powered digital pathology platform
  - Outcome: 30% reduction in review time, 15% accuracy gain
- Light background, rounded rectangles with accent bars.

**Slide 11 — Challenges & Risks**
- Risk matrix or color-coded list:
  - Data Privacy & Security (High)
  - Algorithmic Bias (High)
  - Regulatory Hurdles (Medium)
  - Integration Costs (Medium)
  - Clinician Adoption (Low-Medium)
- Light background, severity colors: red for high, orange for medium, blue for low.

**Slide 12 — Opportunities / Solutions**
- 4 opportunity cards:
  - Personalised Medicine: Genomics + AI-driven treatment plans
  - Early Detection: Cancer screening via computer vision
  - Operational Efficiency: Automated scheduling and billing
  - Global Access: Telemedicine with LLM-powered triage
- Light background, card layout with rounded rectangles.

**Slide 13 — Future Outlook**
- Forecast: "By 2030, 90% of hospitals will deploy clinical AI" and "AI could unlock $1T in annual value across healthcare"
- Supporting bullet points on multimodal AI, agentic workflows, and AI-first biotech.
- Dark background with large white stats.

**Slide 14 — Key Takeaways**
- 5 numbered takeaways with icon circles (navy circles with white numbers):
  1. AI is moving from research to bedside at scale.
  2. Diagnostics and drug discovery lead adoption.
  3. Data governance is the single biggest blocker.
  4. Investment remains strong despite macro headwinds.
  5. The next decade will redefine the clinician–AI relationship.
- Light background.

**Slide 15 — Thank You / Q&A**
- Title: "Thank You"
- Subtitle: "Questions & Discussion"
- Thin orange accent bar at top, dark navy background, centered.

### Technical Constraints
- Single ES module script (`generate.mjs`) using `import pptxgen from "pptxgenjs"`
- No external images, no templates, no base64 — shapes, charts, gradients, text only
- Every slide must include speaker notes (`slide.addNotes(...)`)
- `await pres.writeFile({ fileName: "presentation.pptx" })`
- Must run with `npm install pptxgenjs && node generate.mjs`

### Visual Quality Requirements
- Consistent margins: 0.5" from edges minimum
- Font hierarchy: title 28-36pt bold, subtitle 18-22pt, body 14-16pt, footnotes 9-10pt
- No slide should be just title + bullet list — every slide needs a chart, shape, table, or composed layout
- Use rounded rectangles (`rectRadius`) for cards
- Use navy circles with white numbers as pseudo-icons
- Use thin orange accent bars for visual separation
- Chart data labels enabled, legends positioned cleanly
- Table header: navy fill, white text; alternating rows: `#EDF2F7` / `#FFFFFF`

## Notes

- This prompt is designed to be fed verbatim to any capable LLM to reproduce the exact deck.
- The data is realistic and internally consistent as of 2026.
- To run: `npm install pptxgenjs && node generate.mjs`
- Output: `presentation.pptx`