Clinical Decision-Support Platform — Reliance Jio AI CoE
The strategic problem was clinician trust, not model accuracy.
Reading time: approximately 1 min
The Situation
[FILL: Describe the Jio AI CoE context. What was the product goal and who were the users?]
How I Found the Problem
Ran 100+ user interviews with primary-care doctors in underserved regions. Key finding: doctors would not act on AI recommendations without understanding confidence levels and override options. Trust, not accuracy, was the adoption barrier.
The Real Problem
Clinician trust was the real product problem — not model performance. Doctors needed to feel in control, not replaced.
Options Considered
[FILL: UX and architecture options considered for building trust. How you chose bias gates, confidence thresholds, and manual overrides.]
What We Shipped
[FILL: The trust-first interaction model. Hypothesis, success metric, kill criterion, phased rollout.]
What Moved
500-user beta launched. Government of India partnership initiated for nationwide public-sector rollout.
What I'd Do Differently
[FILL: What would you change in the research or build phase?]