India’s hospital finance system is undergoing a compressed and consequential transformation. Where most mature markets spent two decades incrementally automating their revenue cycles, India is making the leap from largely manual, paper-driven claims workflows to AI-mediated adjudication within a single decade — driven by three converging forces: the Ayushman Bharat Pradhan Mantri Jan Arogya Yojana (AB PM-JAY), which covers the bottom 40% of the population under standardised health packages; the Ayushman Bharat Digital Mission (ABDM), which is building the interoperable digital health infrastructure; and the National Health Claims Exchange (NHCX), designed as the “UPI moment” for health insurance — a single gateway that will allow hospitals, insurers, and third-party administrators (TPAs) to exchange claims data through standardised APIs in near real time. The India Revenue Cycle Management market, valued at approximately USD 5.4 billion in 2025, is projected to reach USD 15 billion by the early 2030s at a CAGR of 12–13%, driven by rising health insurance penetration and the digitisation of claims workflows across both government schemes and private insurance.
On the payer side, AI is already embedded with considerable sophistication. The National Health Authority (NHA) has built one of the most advanced AI-based fraud analytics stacks among low- and middle-income countries. PM-JAY’s AI and machine learning framework analyses massive claims datasets to risk-score hospitals, detect suspect patterns such as impersonation, upcoding, and post-mortem claims, and flag fraudulent transactions in near real time. As of 2023, the system had flagged over 2.15 lakh potentially fraudulent transactions and prevented fraudulent claims worth more than ₹630 crore. The next step — now being piloted through the AB PM-JAY Auto-Adjudication Hackathon, co-sponsored by NHA, IndiaAI Mission, and IISc — moves beyond fraud analytics to near real-time AI-driven adjudication of routine claims. Reported pilots show settlement times collapsing from approximately twenty days to four hours, with AI tools parsing multilingual and low-quality scanned documents, applying standard treatment guideline checks, and auto-deciding claims without human intervention. Private insurers are simultaneously deploying AI for underwriting risk profiling, member enrolment verification, hospital credentialing, and fraud scoring — functions that are publicly framed as efficiency and fraud-control tools but are structurally equivalent to the denial-optimisation engines of US payers.
The denial and repudiation picture in India is already significant, even before AI adjudication fully matures. An IRDAI-linked analysis estimates that approximately 11% of health insurance claims in India were denied in FY24 — amounting to roughly ₹26,000 crore — with primary reasons including documentation errors, non-coverage of pre-existing conditions, and policy lapses. IRDAI data further reveal that health insurers paid only 71.3% of the total ₹1.2 lakh crore in claims filed during FY24, with 36 lakh claims worth over ₹10,000 crore rejected outright and another 20 lakh claims worth approximately ₹7,500 crore remaining pending. These figures precede AI-driven adjudication at scale; as auto-adjudication expands through NHCX, the volume and speed of both approvals and denials will accelerate simultaneously. The critical question India must answer now — while the architecture is still being designed — is whether the algorithms will be built to serve fraud control and patient access, or whether they will drift, as US experience suggests they can, toward systematic denial optimisation that maximises payer margins.
On the hospital side, a new generation of India-specific RCM and claims-automation platforms is beginning to redress the imbalance. Care.fi’sRevNow — launched in late 2024 as an AI-enabled, end-to-end claim management platform — covers the full journey from admission to discharge, pre-authorisation to post-authorisation, and claim submission to settlement. By early 2025, RevNow had processed over ₹800 crore in claims across 300-plus hospitals, supported by fresh debt funding to scale nationally. IHX embeds intelligent OCR, rule-based pre-checks, and AI-assisted validations into hospital billing workflows, with client hospitals reporting 20–25% faster settlement turnaround times, 13–14% reductions in payer query volumes, and nearly 10% falls in repetitive queries. These platforms are India’s answer to US AI-RCM tools like Waystar and RapidClaims: they automate claims hygiene, track pre-authorisations, and accelerate cash flow — though autonomous denial-appeal generation at scale, a feature now standard in the US, is not yet widely deployed in India. The global AI in RCM market, estimated at USD 20.6 billion in 2024 and projected to reach USD 70 billion by 2030, identifies India as among the fastest-growing segments, driven by ABDM adoption, rising insurance penetration, and a proliferation of health-fintech startups building automation layers for claims, coding, and billing.
The equity stakes in India’s AI-mediated revenue cycle are arguably sharper than in any comparable healthcare system. EHR and digital-health adoption is highly uneven: large corporate hospitals — Apollo, Max, Narayana, Fortis — have implemented mature HIS/EMR systems and are positioned to plug directly into NHCX and AI-RCM platforms. But NATHEALTH analyses and academic surveys find that less than 10–20% of rural hospitals have robust digital record-keeping, with most still relying on paper files and non-standardised software. If auto-adjudication and AI-driven compliance checks become the default standard for PM-JAY and private claims — as the trajectory clearly suggests they will — small hospitals and rural nursing homes that cannot invest in interoperable HIS infrastructure will face higher query rates, slower settlements, and more denials due to documentation gaps rather than clinical inappropriateness or fraud. Large hospital chains, conversely, can tune their coding, documentation, and submission processes to the incentive structures of AI models, capturing a disproportionate share of reimbursed revenue. PM-JAY’s own data are instructive: confirmed fraud accounts for only 0.18% of authorised admissions, yet many hospitals — particularly small private facilities in poor districts — face heightened scrutiny or delayed payments because they statistically resemble the risk profile of bad actors. India has a genuine and time-limited opportunity to hard-wire algorithmic accountability into the design of NHCX, auto-adjudication standards, and the Digital Personal Data Protection Act’s health-data rules — mandating transparency, human-review thresholds, contestability mechanisms, and bias auditing before AI adjudication becomes irreversibly entrenched. The policy window is open now. It will not remain so for long.
– Brij Mohan Mandala
