# One-Shot Prompt

**Topic**: AI Applications in Healthcare
**Theme**: Corporate Dark (healthcare adaptation) — deep navy/near-black backgrounds throughout, teal accent
**Generated**: April 2026
**Model**: deepseek-v4-flash

## Prompt

Generate a complete Node.js script using PptxGenJS that produces a professional 15-slide presentation about AI applications in healthcare, using a dark visual theme throughout.

### Theme Specification

All slides use a dark background (`#0A1628`, near-black navy). Use this colour palette:

- Background: `#0A1628` (near-black navy)
- Card background: `#112240` (slightly lighter navy)
- Primary/accent: `#00BFA5` (teal/healthcare green)
- Secondary accent: `#4FC3F7` (light blue)
- Warm accent (emphasis): `#FF6F61` (coral)
- Amber accent: `#FFD54F`
- Text: `#FFFFFF` (white)
- Light/muted text: `#90A4AE`
- Chart colours: `#00BFA5`, `#4FC3F7`, `#FF6F61`, `#FFD54F`, `#B39DDB`, `#F06292`

Use "Arial" as the font throughout (universally available).

### Slide-by-Slide Structure (15 slides, mandatory)

**Slide 1 — Title Slide**: "AI in Healthcare" in 44pt bold white. Subtitle: "Transforming Diagnosis, Drug Discovery, and Patient Care" in 18pt light blue. Decorative oval shapes in the top-right area. Bottom accent bar in teal and navy. Date ("April 2026") and "Generated by deepseek-v4-flash" at bottom.

**Slide 2 — Agenda**: 6 numbered items with teal circle pseudo-icons connected by thin lines. Items: The Healthcare AI Opportunity, Market Landscape, Key Application Categories, Historical Evolution, Traditional vs AI-Assisted Comparison, Case Studies Challenges & Future.

**Slide 3 — Context / Why This Matters**: A highlighted callout box describing healthcare's triple crisis (costs, workforce shortages, aging population) with teal border. Two large-stat cards side by side: "$188B — Projected global AI healthcare market by 2030 (CAGR: 37%)" in teal, and "85% — Of healthcare organisations have an AI strategy (up from 35% in 2020)" in light blue.

**Slide 4 — Key Data Point**: Hero display of "40%" in 80pt teal bold — "Reduction in missed diagnoses when AI augments radiologist review". Three supporting stat cards on the right showing "36M", "2.7B", "47%" with left accent bars in teal, light blue, and coral.

**Slide 5 — Market Landscape (Bar Chart)**: Horizontal bar chart showing AI healthcare market segments: Diagnostics ($45.2B), Drug Discovery ($38.7B), Robotic Surgery ($27.4B), Patient Monitoring ($21.8B), Administrative ($15.3B), Mental Health ($9.6B). Data labels on bars. Use the 6 chart colours.

**Slide 6 — Application Categories (Doughnut Chart)**: Doughnut chart showing deployment share: Diagnostic Imaging (28%), Drug R&D (22%), Genomics (18%), Virtual Assistants (15%), Wearables (12%), Other (5%). Show percentages. Side card with insight annotation.

**Slide 7 — Timeline**: Horizontal timeline line at y=3.0 with 6 milestone nodes (circles) connected by vertical lines. Milestones: 1960s (MYCIN expert system), 1990s (ML in radiology), 2012 (deep learning breakthrough), 2017 (FDA clears first AI diagnostic device), 2020 (AI-accelerated COVID vaccine), 2025 (LLMs in clinical workflows). Summary bar at bottom: "80% of FDA-approved AI devices cleared after 2020".

**Slide 8 — Comparison Table**: 5-row table comparing Traditional vs AI-Assisted across: Diagnostic Accuracy, Drug Discovery Cycle, Cost per Patient, Clinical Trial Success, Administrative Overhead. Each row has the dimension name, traditional value, AI-assisted value (blue), and improvement percentage (teal). Alternating row backgrounds in `#0F1D36` / `#0A1628`.

**Slide 9 — Trend Analysis (Line Chart)**: Multi-line chart with 3 series (Diagnostics, Drug Discovery, Clinical Workflow) across 2019-2026. X-axis: years. Y-axis: adoption percentage (0-90%). Series colours: teal, light blue, amber. Show legend at bottom.

**Slide 10 — Case Study**: Hospital radiology AI deployment — description in a card. Three metric cards below: "94%" (detection rate), "3.2x" (faster workflow), "42%" (fewer false positives). Each metric with a large number, divider line, and label. Bottom banner: "Radiologists reported 3.2x faster workflow and 42% fewer false positives".

**Slide 11 — Challenges & Risks**: 6 challenge items as horizontal cards with left severity indicators (red=coral for High, amber for Medium). Each card: challenge text, severity badge at right. Items: Data Privacy & HIPAA (High), Algorithm Bias (High), Regulatory Uncertainty (Medium), EHR Integration (Medium), Clinical Validation (High), Provider Trust (Medium).

**Slide 12 — Opportunities (Card Grid)**: 2x2 grid of rounded-rectangle cards. Each card: top accent bar in a different colour (teal, blue, amber, purple), icon circle with number, title, description. Cards: "Precision Medicine — Genomic AI tailors treatments", "Remote Patient Monitoring — Wearable AI detects early warning signs", "Drug Discovery Acceleration — Generative AI designs novel molecules", "Clinical Decision Support — Real-time AI assistants reduce errors by 40%+".

**Slide 13 — Future Outlook**: Small line chart (left side, 5.5" wide) forecasting AI Diagnostics and AI Drug Discovery adoption to 2030. Right sidebar with 4 prediction cards showing year and text: 2027 (AI diagnosis standard for radiology), 2028 (first AI-discovered drug in Phase III), 2029 (50% hospitals deploy AI), 2030 ($200B+ annual cost reduction).

**Slide 14 — Key Takeaways**: 5 numbered items with teal circle icons. Each takeaway text in a card: (1) AI transforming every layer of healthcare, (2) Diagnostic accuracy improves 5-10%, (3) Drug discovery timelines collapsing to under 3 years, (4) Regulatory and bias challenges remain critical, (5) Next decade will define equitable delivery.

**Slide 15 — Thank You / Q&A**: Matching title slide styling. "Thank You" in 44pt bold, "Questions & Discussion" subtitle. Decorative ovals, bottom accent bar. Meta info line.

### Technical Requirements

- Single `.mjs` file using ES module imports (`import pptxgen from "pptxgenjs"`)
- No external images — all visuals are shapes, charts, gradients, and text
- Node.js v18+ compatible
- Generate `presentation.pptx` via `await pres.writeFile({ fileName })`
- Speaker notes on every slide (2-3 talking points, transition phrases)
- Footer on slides 2-14 showing topic, date, and page number
- Consistent margins (0.5" minimum from edges)

## Notes

- All data is realistic and internally consistent
- Uses PptxGenJS native chart types: BAR, LINE, DOUGHNUT
- Custom helper functions: `addFooter()`, `addSlideTitle()`, `addCard()`, `addIconCircle()`
- Each slide is wrapped in its own block scope for clean variable isolation
- To run: `npm install pptxgenjs && node generate.mjs`
- Output: `presentation.pptx` (15 slides, dark theme throughout)