The Scale: Roughly 40,000–50,000 Claims a Day
AB PM-JAY processes a very large volume of claims across more than 1,900 treatment packages. NHA/PIB releases during April and May 2026 described the daily volume as roughly 40,000 to 50,000 claims. In April, NHA also said about 15–20% of claims were already being auto-adjudicated, showing that the hackathon was intended to expand an existing capability rather than replace a wholly manual system.
For hospitals, delayed claim settlements mean cash flow crises. For beneficiaries, it means uncertainty about whether their treatment will be reimbursed. For the National Health Authority (NHA), it means navigating a paper-heavy, labour-intensive system where a single claim can take weeks to adjudicate.
In 2026, the government decided enough was enough. It called on India's brightest minds to solve the problem — not with more staff, but with artificial intelligence.
The Government's Call: Build AI to Speed Up Ayushman Bharat Claims
In April 2026, the National Health Authority (NHA), under the Union Ministry of Health and Family Welfare, issued a public notice inviting researchers, innovators, technology professionals, and startups to develop AI-powered solutions for automated claims adjudication under AB PM-JAY. The initiative was called the AB PM-JAY Auto-Adjudication Hackathon 2026.
The hackathon was not a theoretical exercise. It was a direct response to a real operational bottleneck. Manual adjudication leads to delays, inconsistencies, and potential fraud. NHA wanted scalable, technology-driven approaches that could improve turnaround time while maintaining accuracy, transparency, and regulatory compliance.
Registrations closed on 13 April 2026. A PIB release that day said more than 2,600 participants had registered at that point. The later showcase focused on the quality and deployability of solutions rather than publishing a final registration total.
The Grand Finale at IISc Bengaluru
The hackathon culminated in a two-day offline showcase at the Indian Institute of Science (IISc), Bengaluru, on May 8 and 9, 2026. The event brought together policymakers, healthcare leaders, technology experts, insurers, Third Party Administrators (TPAs), and AI startups under one roof.
Senior officials who graced the inaugural session included Shri S. Krishnan, Secretary, Ministry of Electronics and Information Technology, who delivered the keynote address. Dr. Sunil Kumar Barnwal, CEO of the National Health Authority, Prof. Govindan Rangarajan, Director of IISc Bengaluru, and Ms. Jyoti Yadav, Joint Secretary (PMJAY) at NHA, also addressed the gathering. Healthcare industry leader Dr. Devi Prasad Shetty was scheduled to participate.
The event was organised in collaboration with the IndiaAI Mission, Digital India, and MyGov India — underscoring the government's commitment to building AI solutions on India's digital public infrastructure.
Three Problem Statements: Where AI Meets Healthcare Claims
The hackathon focused on three critical problem areas where AI could transform claims adjudication:
1. Clinical Document Classification and Compliance with Standard Treatment Guidelines (STGs)
Winning teams showcased AI solutions capable of automatically classifying healthcare documents and performing multilingual OCR on low-quality scans to extract structured clinical and billing information. These systems could identify mandatory visual elements such as stamps and signatures, and generate explainable adjudication outputs based on Standard Treatment Guidelines and policy compliance requirements.
2. Radiological Image-Based Condition Detection and Report Correlation
Teams demonstrated assistive AI solutions that could interpret radiological images such as X-rays, CT scans, and MRIs. These tools help adjudicators understand image contents, correlate findings with hospital-submitted reports, and validate the claimed condition, disease stage, and treatment timeline under Standard Treatment Guidelines.
3. Document Forgery and Deepfake Detection
Perhaps the most critical problem statement focused on fraud prevention. Winning teams presented AI/ML-based solutions aimed at detecting forged medical documents submitted during claims processing — including tampered discharge summaries, manipulated bills, ghost identities, and fabricated medical reports. These solutions were designed to support digital claims adjudication and strengthen programme integrity under AB PM-JAY.
What the Hackathon Demonstrated — and What It Did Not
The official NHA/PIB releases emphasised reducing manual effort, improving speed, expanding auto-adjudication and strengthening fraud detection. They did not establish a verified scheme-wide result that average processing time had already fallen from 20 days to four hours, so that claim should not be presented as an NHA outcome.
The practical takeaway is narrower: NHA was seeking to expand auto-adjudication so that more claims could be processed faster and with less manual intervention, while improving consistency and fraud detection. The official releases cited here do not quantify a scheme-wide reduction to a specific number of hours.
The hackathon aimed to expand these capabilities across a wider range of medical packages and specialties, bringing the benefits of AI-driven claims management to more hospitals and beneficiaries across India.
Why This Matters: The Scale of PM-JAY Fraud and Inefficiency
To understand why AI-driven claims adjudication is urgent, one must understand the scale of the challenge. AB PM-JAY covers 50 crore+ beneficiaries across 28,000+ empanelled hospitals. At that scale, fraud, waste, and abuse are structural inevitabilities.
The NHA has identified fraud as the primary operational challenge for the scheme's continued viability. Common fraud patterns include:
- Ghost patient claims: Claims submitted for patients who were never hospitalised at all.
- Phantom provider billing: Claims from facilities that performed no services.
- Upcoding: Billing for procedures not performed or of higher complexity than actually delivered.
- Empanelment fraud: Facilities obtaining PM-JAY empanelment through falsified credentials.
- Document forgery: Tampered discharge summaries, manipulated bills, and fabricated medical reports.
Traditional manual auditing cannot keep pace with these fraud patterns. AI can. Machine learning models can analyse millions of claims in the time it takes a human auditor to review a handful. They can detect cross-hospital fraud patterns that are invisible when each claim is reviewed in isolation.
The ABDM Advantage: India's Unique Data Infrastructure
ABDM and ABHA provide important digital-health infrastructure, but an ABHA number should not be described as automatically creating a complete longitudinal health record available for claims. Health-data exchange remains consent-based and depends on participating systems, data availability and applicable privacy controls.
This data infrastructure is a game-changer for fraud detection. A patient who receives the same procedure at three different empanelled hospitals in the same month, billed separately to PM-JAY each time, becomes visible as a fraud pattern at the national level. Without ABDM data exchange, that pattern is invisible, because each hospital's claim looks legitimate in isolation.
Cross-provider fraud detection using ABDM-linked health records, with explicit patient consent via the Health Information Exchange and Consent Manager (HIE-CM) and in compliance with the Digital Personal Data Protection Act 2023, is the most powerful fraud prevention capability available to PM-JAY.
Panel Discussions: The Future of AI in Indian Healthcare
The hackathon showcase featured thought-provoking panel discussions bringing together leading voices from government, industry, startups, and academia.
A session on "Building AI for Indian Healthcare" explored strategies for scaling AI innovations within the healthcare ecosystem, leveraging the robust digital public infrastructure established under ABDM. Panellists highlighted the importance of workflow integration, validation frameworks, quality datasets, edge deployment, privacy safeguards, and scalable implementation pathways.
Another panel on "Future of Claims Adjudication" deliberated on how AB PM-JAY can enable a faster, more transparent, and accountable claims ecosystem through the integration of artificial intelligence, automation, interoperable digital platforms, and data-driven governance frameworks.
A dedicated session on "Fraud, Waste and Abuse in the Era of AI — Challenges and Opportunities" examined the responsible use of AI in fraud detection, while addressing critical issues related to transparency, accountability, and data privacy.
Regulatory Guardrails: AI Can Approve, AI Can Flag, But AI Cannot Deny
The hackathon discussions stressed responsible AI, validation, transparency, accountability and human workflow integration. The official releases do not establish a binding rule that AI may approve claims but can never autonomously deny them; any such operational safeguard should be sourced to the specific PM-JAY process or regulation before being stated as mandatory.
For the parallel regulatory discussion in insurance, see our IRDAI AI governance working-group explainer, covering transparency, explainability, claims and fraud-prevention safeguards.
Clinical coverage determinations, fraud-based claim denials, and prior authorization denials all require human review before any communication reaches the member or provider. This is not an engineering constraint. It is a patient protection requirement built into insurance regulations.
The discussions also focused on the role of Small Language Models (SLMs), Large Language Models (LLMs), and multimodal AI systems in addressing diverse healthcare use cases, particularly in low-resource and local language settings. India is uniquely positioned to demonstrate how AI can build upon the country's digital public infrastructure to deliver services at scale across sectors such as healthcare, education, and agriculture.
What This Means for Private Health Insurance in India
Private health insurers also use automation and analytics for claims and fraud management, but this PM-JAY hackathon did not establish a current IRDAI estimate that private-health-insurance fraud costs Rs 45,000–60,000 crore annually. That figure should not be attributed to IRDAI without a direct source.
Private insurers, including Star Health, HDFC ERGO, New India Assurance, and ICICI Lombard, are actively deploying AI claims processing and fraud detection systems. The TPA ecosystem is the integration layer for most of these deployments, with cashless authorisation workflows being the highest-value fraud prevention touchpoint.
The IRDAI Insurance Fraud Monitoring Framework Guidelines 2024 mandate that all insurers maintain systematic fraud detection and prevention processes. Insurers running rule-based fraud detection against the IRDAI framework are operating at the compliance floor. Those deploying AI are building a durable competitive advantage.
What This Means for Policyholders and Hospitals
For beneficiaries, AI-driven claims adjudication means faster settlements, fewer delays, and greater transparency. When a claim is processed in hours instead of weeks, hospitals get paid faster, and patients face less uncertainty about their coverage.
For hospitals, particularly smaller empanelled facilities that struggle with cash flow due to delayed reimbursements, AI adjudication is a lifeline. Faster payments mean better working capital, more reliable operations, and greater willingness to participate in the PM-JAY scheme.
For the healthcare ecosystem as a whole, AI-driven fraud detection means that genuine claims are processed more swiftly, while fraudulent claims are caught before payment is released. This protects the scheme's financial sustainability and ensures that resources reach those who genuinely need them.
The Road Ahead: From Hackathon to Real-World Deployment
The AB PM-JAY Auto-Adjudication Hackathon 2026 was not the end of the journey. It was the beginning. The winning solutions exhibited strong potential for real-world implementation, and NHA has signalled its intent to integrate the most promising innovations into the existing AB-PMJAY infrastructure.
The next steps involve:
- Piloting winning solutions on live claims data.
- Integrating AI tools with existing hospital management information systems (HMIS) and claims portals.
- Building validation frameworks to ensure AI decisions are explainable, auditable, and fair.
- Training adjudicators to work alongside AI tools rather than being replaced by them.
- Expanding auto-adjudication capabilities to cover more treatment packages and specialties.
Bottom Line
The AB PM-JAY Auto-Adjudication Hackathon 2026 showcased AI solutions for document classification, radiology-based validation and forgery/deepfake detection. NHA said 15–20% of claims were already auto-adjudicated in April and described a daily claims volume of roughly 40,000–50,000 across 1,900+ packages. Selected winning solutions may be considered for future deployment; the event itself did not prove a universal 20-day-to-four-hour processing reduction.
For India's 50 crore PM-JAY beneficiaries, this means faster claims, fewer delays, and greater trust in the system. For hospitals, it means quicker reimbursements and better cash flow. For the healthcare ecosystem, it means a more efficient, transparent, and fraud-resistant claims process.
At Tatkal Claims, we believe technology and advocacy go hand in hand. While AI is transforming claims processing at the systemic level, individual policyholders still face arbitrary rejections, delays, and underpayments every day. If your health insurance claim has been rejected, delayed, or underpaid — whether under PM-JAY or a private policy — our legal team is here to help you fight for the settlement you deserve.
Facing a delayed or rejected health insurance claim? Contact our legal team at Tatkal Claims for expert assistance in challenging unfair denials and securing the settlement your family deserves.
