Rebuilding the Group-Health Claims Desk Around NHCX: A Broker Operations Guide for 2026

For most Indian broking houses, the corporate group-health claims desk grew up as an email-and-portal operation. A member or HR raises a query, the desk forwards it to the TPA, the TPA replies on its own portal, and the broker stitches the two together by hand. That model assumed the broker sat outside the data flow and added value through chasing and relationship pressure.

The National Health Claims Exchange (NHCX), built by the National Health Authority with IRDAI, breaks that assumption. NHCX is a switch that carries cashless pre-authorisation requests, queries, and claim adjudication messages between hospitals, insurers, and TPAs over a single FHIR-based protocol. The goal is to replace dozens of bilateral hospital-to-insurer portals with one standard exchange. Since the 2024 rollout, the large majority of health insurers and TPAs, covering most of the retail and group health market, have connected to the sandbox or live environment, and 2026 phases are onboarding smaller insurers and more hospitals.

Two regulatory clocks now sit on top of this pipe. Cashless authorisation requests must be decided within a defined window (one hour on receipt of the request), and final discharge approval is expected within three hours of the hospital sending the discharge bill. Separately, IRDAI's tightened master-circular regime, reinforced through 2025 and 2026, pushes insurers to settle the large majority of genuine claims within fifteen calendar days of the final document, with interest consequences for delay.

The practical message for broking leadership is blunt. If your claims desk SOP was written before NHCX, it is now describing a workflow that no longer matches how the cashless decision actually travels.

To redesign the desk, you need a clear picture of what moves where. Under the legacy model, a hospital logged into an insurer or TPA portal, keyed in patient and treatment details, and uploaded documents. Each insurer ran a different portal with different fields. The broker saw none of this directly and relied on the TPA to narrate status.

Under NHCX, the same interaction becomes a set of structured FHIR resources passing through the exchange. A cashless request carries the patient identity, policy or member reference, diagnosis and procedure codes, estimated cost, and supporting documents as defined data objects rather than free-text uploads. The insurer or TPA responds with a coded decision: approved, queried, or denied, each with a reason. Because the message is standardised, every participant is reading the same fields.

Three consequences matter for the desk.

1. Field discipline becomes a claims issue, not just an IT issue. If the hospital sends a procedure code that does not map cleanly to the policy benefit, the query bounces back faster, but it also bounces back in a form the broker can diagnose. Mismatched fields, not missing relationships, will drive a growing share of authorisation friction.
2. Timestamps are now adjudicable. Each message hop is logged. When HR asks why an approval took six hours, the answer is no longer anecdotal. The desk can point to the exact leg where the clock was held.
3. The TPA stops being the only source of truth. Where the broker has read access to exchange-level status, the desk can verify the TPA's narrative against the actual message log rather than accepting it.

None of this happens automatically. NHCX defines the pipe; it does not redesign your desk. The broker that treats NHCX as a passive infrastructure upgrade will see the same service complaints. The broker that instruments around it gains a defensible audit trail.

The three-hour discharge clock is the single most demanding operational constraint NHCX puts on a group-health programme. A member is medically ready to leave; the family is waiting; the hospital has sent the final bill. From that moment, the regulatory expectation is that the insurer decides within roughly three hours. Every minute beyond that is a service failure that HR feels and the broker owns.

A desk built for this clock looks different from a desk built for email. Design around four principles.

Pre-stage the predictable cases

Most planned admissions are visible days ahead. The desk should pre-verify member eligibility, sum insured availability, sub-limits, and any waiting-period exposure before admission, not at discharge. If the hard questions are answered on day one, the discharge message becomes a confirmation rather than a fresh adjudication.

Triage by clock risk, not by arrival order

Not every case carries the same clock risk. A clean planned procedure inside network with a comfortable sum insured is low risk. An emergency admission with a borderline diagnosis-to-benefit mapping is high risk. Route the high-risk cases to senior handlers immediately rather than processing in the order they land.

Hold the TPA to message-level timestamps

With NHCX, the desk can ask a sharper question than "where is the approval". It can ask "which leg is holding the clock right now". Build your escalation script around that.

Keep a live discharge board

A shared, real-time view of every open discharge against its clock lets the desk escalate before the window closes, not after HR calls. This is the operational heart of the redesign: the desk watches clocks, not inboxes.

The shift is cultural as much as technical. A reactive desk answers queries. A clock-aware desk manages a countdown.

Cashless is only part of the picture. A meaningful share of group-health claims still flow as reimbursement, especially for non-network treatment, certain emergencies, and benefits that sit outside the cashless arrangement. For these, the binding constraint is the settlement clock rather than the authorisation clock.

IRDAI's reinforced regime expects insurers to settle the large majority of genuine claims within fifteen calendar days of receiving the final document set, with interest payable when they run past the window. The operative words are final document set. The clock that matters starts when the file is complete, which means the broker desk's real lever is document completeness, not chasing speed after submission.

This reframes the reimbursement workflow.

• Run a completeness gate at intake. Before a reimbursement file leaves your desk, it should pass a checklist mapped to the specific insurer's documentation list: discharge summary, final bill with break-up, payment proof, investigation reports, and the policy-specific declarations. A file submitted incomplete simply has not started its clock.
• Date-stamp the complete-file moment. Record when you believe the file became complete and submit that view to the insurer. If a dispute arises later about interest, your contemporaneous record of completeness is the strongest evidence you can hold.
• Separate genuine delay from genuine query. A legitimate query that asks for a missing document resets the practical clock. A query that re-asks for something already supplied does not, and the desk should push back on it in writing.

For large corporate programmes, the desk should report monthly on the proportion of reimbursement claims settled inside fifteen days and the value of any interest that became due. That single metric tells the client more about insurer behaviour than a settlement ratio ever will.

NHCX does not remove the TPA from group-health; for most corporate programmes the TPA still adjudicates and services. What changes is the basis on which the broker holds the TPA to account. Governance built on portal screenshots and monthly MIS becomes governance built on message-level evidence.

Revisit the service-level construct in the broker-TPA arrangement. The traditional service standards (turnaround on authorisation, query resolution, settlement) were always asserted and rarely provable. With exchange timestamps, they become measurable. Rewrite the standards so they reference the events NHCX actually logs: time from request receipt to decision, time spent in query, number of repeat queries on the same case.

Three governance moves follow.

First, demand exchange-aligned reporting. Ask the TPA to report performance against the same timestamps the exchange records, not against its internal portal clock. Where the two diverge, the divergence is itself a finding.

Second, track repeat-query rate as a quality signal. A TPA that resolves cleanly on the first pass is performing; one that generates multiple queries per case is either under-resourced or using queries to stop the clock. The exchange makes this visible in a way the old portals did not.

Third, separate the TPA's performance from the insurer's. On a cashless decision, the request may pass through the TPA to the insurer and back. When a case runs slow, the desk needs to know which party held it. Build that attribution into your monthly governance pack so the client conversation is precise rather than a general complaint about "the TPA".

For board-level oversight of TPA arrangements, the deeper playbook in our TPA governance for commercial health piece still applies; NHCX simply gives that governance harder evidence. The principle holds: you cannot govern what you cannot timestamp, and now you can timestamp almost everything.

A claims desk that reads structured health data carries obligations the old forwarding desk could pretend to ignore. NHCX moves diagnosis codes, procedure codes, and clinical documents in machine-readable form. That is sensitive personal data, and the Digital Personal Data Protection regime applies to how the broker handles it.

The desk should treat three points as non-negotiable.

Consent must cover the actual flow. A member's consent for the broker to assist with their claim should contemplate that the broker desk may view structured clinical data passing through the exchange or shared by the TPA. Generic enrolment consent written before NHCX may not describe this clearly. Review the consent language in your group-health onboarding against the data the desk now actually touches.

Minimise what the desk retains. The temptation, once structured data is available, is to warehouse everything for analytics. Resist it at the claims-desk layer. The desk needs enough to manage the clock and resolve the case; it does not need to become a clinical data lake. Retention and access controls should reflect the sensitivity of the category.

Keep the audit trail clean of unnecessary clinical detail. When the desk logs a case for governance, log the operational facts (timestamps, decision codes, query reasons) rather than copying diagnostic narrative into general-purpose trackers that a wider team can read.

The same exchange that gives the broker a stronger audit position also raises the stakes on data handling. A desk that stores structured health data loosely converts an operational upgrade into a compliance exposure. Decide your retention and access policy before you switch on exchange-level visibility, not after.

For the phased compliance timeline that frames these obligations, see our analysis of the DPDP rules and the insurance compliance clock. The practical takeaway: the claims desk is now a data controller in everything but name, and should behave like one.

Most broking houses cannot rebuild the desk overnight, and they should not try. A staged plan over a quarter gets you to a defensible NHCX-aligned operation without breaking live service.

Weeks 1 to 3: map the current flow. Document exactly how a cashless case and a reimbursement case move today, including every email, portal, and handoff. You cannot redesign what you have not mapped. Identify where the desk currently has no visibility and where it relies entirely on TPA narration.

Weeks 4 to 6: secure exchange-aligned visibility. Work with your insurer and TPA partners to understand what NHCX-level status you can see, directly or through their reporting. Where direct visibility is limited, negotiate timestamp-aligned reporting into the service arrangement. Define the named events you will track: request received, decision issued, query raised, file complete.

Weeks 7 to 9: rebuild the SOP around clocks. Rewrite the desk SOP so it organises around the three-hour discharge clock and the fifteen-day settlement clock rather than around inbound query type. Introduce clock-risk triage, the live discharge board, and the completeness gate for reimbursement.

Weeks 10 to 12: instrument governance and train. Stand up the monthly governance pack that attributes delay to the correct party and reports settlement-within-window and interest-due metrics. Retrain handlers from a query-answering mindset to a clock-managing one, and test the escalation script against real cases.

Throughout, keep the client HR team informed. The single biggest reputational win available to a broker in 2026 is to walk into the renewal review with timestamped evidence of how the programme actually performed against the regulatory clocks. That is a conversation the old screenshot-driven desk could never have. The houses that rebuild now will set the service standard the rest of the market gets measured against.