India crossed 50,000 NQAS certifications in 2025, but 85% of facilities remain uncertified. The gap isn't effort — it's systems. Here's how AI is becoming the missing infrastructure layer in public health quality.
India hit a landmark in January 2026 — 50,000 public health facilities certified under the National Quality Assurance Standards. That number sounds impressive until you look at what it actually means.
There are roughly 1.84 lakh Ayushman Arogya Mandirs alone. Add district hospitals, sub-district hospitals, community health centres, and urban health centres — the total public health infrastructure crosses 2 lakh facilities. At 50,000 certifications, we've covered roughly a quarter.
The government's own interim target is 50% certification by March 2026. The math doesn't work. And the deadline doesn't care.
But this piece isn't about the deadline. It's about understanding why quality keeps failing in Indian public health despite decades of effort — and why artificial intelligence isn't just a nice-to-have anymore. It's the missing infrastructure layer.
The Inconvenient Truth About Quality in Indian Healthcare
Here's what the numbers don't show you.
India loses an estimated 16 lakh lives every year to poor quality of care. Not because healthcare doesn't exist — because the quality of care inside those facilities doesn't meet basic standards. A Lancet Global Health Commission study put it starkly: more people in low- and middle-income countries die from poor quality care than from lack of access to care.
At the district level, the cost of poor quality is devastatingly tangible. Our field assessments across districts in Chhattisgarh, Jharkhand, and Bihar reveal a consistent pattern: every district hemorrhages between ₹6 to 8 crore annually in quality-related losses. This includes hospital-acquired infections, avoidable repeat OPD visits and duplicate diagnostics, medicine and consumable wastage, legal costs from FIRs and patient compensation, equipment breakdown, and staff productivity loss from poor SOP adherence.
These aren't projections. These are observed realities. The money is being lost right now, in every district, every year.
Why Conventional Quality Programmes Keep Failing
The NQAS framework itself is robust. It's built on global best practices, aligned with International Society for Quality in Healthcare standards, and recognized by IRDAI for hospital empanelment. The framework isn't the problem.
The execution model is.
Here's what typically happens when a district receives an NQAS mandate: A district quality team — usually 10 to 15 people already carrying full workloads — is told to prepare 150+ facilities for certification. The mandate requires 500+ checkpoints per facility, SOP creation across departments, documentation systems, training, and eventually a national assessment. The load on an already saturated system jumps from 100% to 150%.
The result is predictable. Teams choose between running daily clinical operations and preparing for quality certification. They can't do both. Urgency wins. Quality gets pushed to "next quarter." Certification stays pending. Year after year.
A peer-reviewed study from Bhavnagar, Gujarat — one of India's positive deviant districts for NQAS — found exactly this tension. Facility staff reported that preparing for NQAS disrupted daily routine programmes. The researchers noted that the certification process itself often took precedence over deeper meaningful changes in quality of care.
This is the zero-sum trap. The system doesn't have bandwidth for quality AND operations simultaneously. And since nobody has solved the bandwidth problem, quality keeps losing.
The Pattern Nobody Was Connecting
We discovered this pattern firsthand.
In 2024, Jaimini Group arrived in Sahibganj, Jharkhand to establish functional delivery wards at Ayushman Arogya Mandirs under Janani Jeevan Jyoti, a maternal health initiative. The task was straightforward: assess centres, build wards, improve maternal health outcomes.
But the moment we started visiting facilities — talking to CHOs, reviewing registers, walking through labour rooms — we noticed something we weren't looking for.
At the district hospital, the same problems. At the CHC, the same issues. At every PHC and Ayushman Arogya Mandir we visited — the same invisible weight pulling everything down. NQAS. The foundational quality layer underneath every national health programme. And it was failing everywhere.
The insight was simple but profound: every health programme — maternal health, disease elimination, NCD screening, immunisation, ABDM — runs on the same CHO, at the same facility, within the same 8 hours. NQAS is the base layer. If it fails, everything built on top eventually collapses.
Nobody was connecting these dots because every programme was being treated as a separate vertical. But on the ground, they all converge on the same person.
Why AI Is Not Optional Anymore
The conventional approach to quality implementation is fundamentally memory-based. A training happens in a classroom. Information goes into someone's head. They return to their facility. Within weeks, the knowledge degrades. The facility reverts.
This model worked (barely) when you needed to certify 10 district hospitals. It breaks completely when you need to certify 1,500 facilities across a state — each with different gaps, different team capabilities, different infrastructure levels — within 12 months.
This is where AI stops being a technology trend and becomes an execution necessity.
AI addresses the three fundamental bottlenecks that have kept quality implementation stuck for a decade:
The training bottleneck.
You cannot conduct in-person training for 5,000 health workers across 150 facilities simultaneously. But you can deploy an AI-powered learning platform that delivers NQAS training in video, audio, and interactive formats — in local languages, at each worker's pace, with competency tracking that shows exactly who has learned what.
The assessment bottleneck.
You cannot manually assess and track 500+ checkpoints across 150 facilities in real time. But you can build an AI system that digitises gap assessment, auto-generates facility-specific action plans, and provides live readiness scores that update as improvements happen.
The interview bottleneck.
Most facilities that fail NQAS don't fail on infrastructure. They fail at the assessment interview stage. AI-powered mock interviews — conducted repeatedly, at scale, simulating actual NQAS assessment scenarios — eliminate this failure point entirely.
What We Built — And Why It Works
SAAVI ONE — Smart Learning for Health Workers
SAAVI ONE is an AI-powered training and monitoring platform designed specifically for Indian public health workers. It converts dense, 100 to 500 page NQAS standards documents into digestible learning content — videos, audio modules, infographics, and situation-based scenarios. It supports multiple Indian languages. It works in offline mode for facilities in remote areas.
The real innovation is the monitoring layer. Every health worker's progress is individually tracked. District leadership gets a competency dashboard showing who has completed training, who is lagging, who are the top performers, and where intervention is needed. For the first time, health worker capability becomes data-driven — not assumed.
NQAS AI — The Quality Intelligence Engine
NQAS AI transforms quality implementation from a memory-based process to a system-driven one. It handles digital gap assessment — structured, scored, and auto-prioritised. It generates facility-specific action plans with timelines, responsibilities, and resource requirements. It provides live readiness scoring that updates in real time.
The game-changing feature: AI-powered mock interview rounds. Before the national external assessment team visits, every facility team goes through multiple rounds of AI-simulated interviews that mirror actual NQAS assessment scenarios. By the time actual assessors arrive, the surprise factor is eliminated.
AI for CHO — The Missing Layer
When we deployed SAAVI ONE and NQAS AI in the field, we hit an unexpected wall. Not a technology wall — a literacy wall. Community Health Officers had smartphones and capability, but they'd never used an AI tool.
Before you deploy AI platforms, you need to build AI infrastructure inside the person.
So we created "AI for CHO" — a structured, hands-on, one-day training programme that makes every CHO AI-literate using freely available tools. In our first deployment in Kabirdham, Chhattisgarh, we trained 90 CHOs and achieved 93% AI readiness. Each CHO reported saving approximately 15 hours per month. The strongest proof? The CHOs demanded Part 2.
The Model: Project Pravidhi
All of this — the expert consulting team, the AI platforms, the training methodology, the field-tested processes — comes together in Project Pravidhi, our district-level health transformation programme.
Project Pravidhi isn't a consulting engagement where someone writes a report and leaves. It's a full end-to-end ownership model. We deploy senior NQAS experts on the ground — each with 400+ facility certifications — supported by SAAVI ONE and NQAS AI. From initial gap analysis to final national certification, one partner takes complete responsibility.
The model has been designed and piloted across three states:
- Chhattisgarh (Kabirdham): India's first AI Future-Ready CHO District — 90 CHOs trained, AI literacy foundation established.
- Jharkhand (Sahibganj): Origin district — where the pattern was discovered and Project Pravidhi was conceived.
- Bihar (Vaishali): Pilot validated — health workers engaged with genuine energy for the first time. Belief came back.
- Jharkhand (Deoghar): Currently deploying — full 10-centre NQAS transformation funded under DMFT.
The Multiplier Effect
When a district's facilities achieve NQAS certification, it doesn't just tick one box. It creates a multiplier across the entire health ecosystem:
- NHM mandates — maternal health, immunisation, and child health targets get met systematically because quality-certified facilities have the SOPs, training, and systems to deliver consistently.
- Disease elimination — TB, Malaria, and Leprosy programmes run better on facilities with infection control SOPs, trained teams, and AI-assisted documentation.
- ABDM and Digital India — AI-literate CHOs adopt ABHA, e-Sanjeevani, and NCD portals faster because the technology resistance barrier has been broken.
- Kayakalp, LaQshya, MusQan — these quality-adjacent national awards become automatically achievable because NQAS compliance covers most of their checkpoints.
The financial case: districts save ₹2 crore+ annually in quality losses while generating ₹1.5 to 2.25 crore in direct returns over three years. Against under ₹10 lakh per facility investment, the payback period is approximately 6 months.
What Needs to Change
India's NQAS framework is sound. The government's commitment — evidenced by the 50,000 certification milestone — is real. The funding mechanisms, including DMFT for mining-affected districts, exist. What's missing is the execution layer.
Three shifts would accelerate India's quality transformation dramatically:
- From memory-based to system-driven implementation. Quality processes need to live in systems, not in people's heads. AI platforms make this possible at scale.
- From training-first to AI-literacy-first. Deploying digital health tools on a workforce that has never used AI is like installing software without an operating system. AI literacy must precede platform deployment.
- From certification-driven to sustainability-driven. When teams chase patient outcomes instead of certification scores, quality becomes self-sustaining.
India has the infrastructure. It has the people. It has the political will. What it needs now is the execution intelligence to connect them all.
“That's the layer we're building.”
— Jaimini Group Editorial Team
Jaimini Group is a Healthcare Intelligence & Execution Company operating at the intersection of human expertise and AI. Through Project Pravidhi, SAAVI ONE, and NQAS AI, we deliver end-to-end quality transformation for public health districts. Currently active in Chhattisgarh, Jharkhand, and Bihar.