who-03-ideal-rate

AI as Healthcare Levels: A Framework for WHO-India

The Stack (Bottom-to-Top)

Level	Care Type	AI Type	Function	Response Time	Market Signal
⚪ 5	Inpatient	Embodied AI	Integrates all layers into adaptive, physical systems	Aperiodic / On Discharge	Emerging
🟡 4	Outpatient	Generative AI	Creates content, plans scenarios, explores possibilities	Weeks - Month	$50B+ ARR (2025)
🟢 3	Urgent Care	Agentic AI	Executes workflows, automates routine decisions	Days - Week	$100B ARR (2029)
🔵 2	Emergency	Perception AI	Detects patterns, senses changes, triggers alerts	Minutes - Hour	Foundation
🔴 1	First Aid	World AI	Contextual understanding of environment/domain	Seconds - Minutes	Infrastructure

---

Reading the Stack

Top-Down (Integration → Foundation)

5. Inpatient / Embodied AI coordinates physical-digital integration in complex scenarios
4. Outpatient / Generative AI supports planning and creative problem-solving
3. Urgent Care / Agentic AI responds with autonomous workflows ← explosive growth zone
2. Emergency / Perception AI monitors for signals requiring attention
1. First Aid / World AI provides baseline context (maps, sensors, knowledge graphs)

Why Agentic in Middle?

• OpenAI projects Agentic AI = $100B ARR by 2029 (their largest growth segment)
• Agents require reliable perception infrastructure to function
• Like healthcare: urgent care coordinates between emergency triage and outpatient planning
• Foundation enables scale; the action layer captures value

---

For WHO-India Context

Vocabulary Expansion: Instead of asking “should we use AI?”, ask:

• Do we need Embodied AI for robotic surgery at scale?
• Do we need Generative AI to create training materials?
• Do we need Agentic AI to automate vaccine logistics? ← highest ROI zone
• Do we need Perception AI to detect disease outbreaks?
• Do we need World AI to map health system context?

Investment Signal: The market is betting on the middle layer (agents doing work), but it requires the bottom layers to function reliably.