21 April 2026
Three things have become true at once. Jio Allianz is months from a launch date that gives us a captive payer, go-to-market, and need to now manage this book of business. The largest block of private insurance claims in India is cardiometabolic disease and its hospital-borne complications, most of which is preventable if one can see the trajectory. Second, the unit cost of deep phenotyping is falling quick enough to make population-scale biobanking economic for the first time in any market. One implication of this is that we can build the first population scale real world evidence base of biological and medical data. Third, the quality of clinical AI inference is approaching that of highly trained human clinicians while the cost is collapsing. The ability of AI to read, interpret and convert this base of longitudinal data to completely change how health, medicine, and care gets delivered is a once in a lifetime moment. The window to take the lead in defining the next generation health and clinical model in India is the next twenty-four months. This memo proposes what to build, why only we can build it, and what to decide first.
What exists. Primary care in India is not scarce. Pharmacies, clinics (licensed and unlicensed), diagnostic chains, and telehealth saturate every pincode. Patients route themselves to whoever is closest, cheapest and fastest when symptomatic, often too late, and they are largely satisfied with the immediate relief they receive, even when it is of poor clinical quality. Corporate clinic chains have struggled with unit profitability. Adding more branded primary care capacity does not guarantee usage nor change outcomes, does not shift claims, and does not build a franchise.
What doesn’t. Two things break in Indian private health, and both are addressable at our scale. The first is that the relationship between a member and a health system is transactional and episodic. People meet their clinician in the crisis, not before. The second is that secondary and tertiary care are navigated alone. Indian families facing a cardiac scare, a cancer diagnosis, a recommended surgery, or a new chronic diagnosis enter a system that over-tests, steers to higher-margin procedures, charges opaquely, and releases them without follow-up. This is where private insurance claims leak, and this is where the human cost concentrates.
What we are building. A health operating system. The organizing idea is not a clinic chain and not just a new insurance product with incremental benefits. Neither of those will change the market. There are 5 components to our model: 1) Insurance Base which provides underwriting and a GTM strategy 2) Proactive Engagement Stack 3) Navigation Layer 4) Data Architecture and Biobanking Layer 5) Medical and Scientific Asset Layer. AI is central to each layer in our model. Our model develops a trusted relationship with the member in their healthy years, which then proves its value to the member when they need to navigate secondary and tertiary care,
and compounds across their lifetime into the most valuable longitudinal health cohort and dataset in the world.
How are we building it. The guiding design principle is to focus on the levers that give compounding unit value. Revenue, clinic footprints, and premium products can all be copied by any well-capitalized competitor. One asset cannot be copied at our scale within a decade: patient-years of consented, longitudinal, multi-omic phenotype linked to navigation decisions and outcomes. Every architectural choice that follows is in service of compounding this asset faster than any other actor in India, or in fact globally outside the Nordic and UK biobank consortia. Trust is the entry. Data is the flywheel. Outcomes are the proof.
Primary care lives inside the operating system but is not the product. Since primary care is already supplied by the Indian market at density and price, our primary care presence is the Personal Agent, the CHW cadre, and a thin physical footprint where in-person touch is clinically material. Physical infrastructure is data capture and trust maintenance. It is not a retail business and will not be sized as one.
The social contract. The founding story of this company was trust offered by ordinary Indians in return for industrial dividends. The new story asks the same Indians to trust us with their health data and their navigation choices in return for better outcomes and dividends from the bio discovery. The same civilizational contract, in a new asset class. A pharmaceutical firm cannot credibly offer this, because it is a buyer of the data. A hospital chain cannot offer this, because it is a seller of procedures. A foreign actor cannot offer this, because it cannot promise return of value. Only we can.
The Operating System
1) Insurance Base which provides underwriting: This provides us a GTM with a recognizable product type/ wallet share in the market, minimizing customer education and slow product awareness. Jio Allianz is the anchor buyer but any Indian group insurer can be a sufficient proof vehicle in the interim if needed. This is also the feedback loop which disciplines the care side by the insurer’s profit and loss. Every clinical action, from a remote consultation to a recommended procedure, runs against a payback loop. Interventions that do not demonstrably reduce claims are retired. This is how care becomes economically rational rather than volume driven.
2) Proactive Engagement Stack: This is the surface the member sees every day. A Personal Agent lives on the member’s device and holds their personally identifiable information locally. A Clinical AI inference layer processes consented data on demand and retains nothing; it is stateless by design, so the compute layer can never become a honeypot. A Tech-Powered and Jio-Trained Community Health Worker cadre at a density of one per thousand members
provides the human escalation path when the agent judges that human contact is needed. This is where trust is formed.
3) Navigation Layer: This is the layer that activates when a clinical signal appears or an event looms. A human navigator, drawn from the CHW cadre and a specialist coordinator team, walks the member through their primary, secondary or tertiary care. Four contract tiers substitute for hospital and clinic ownership. By not owning or deploying large capex into fixed structures, our model gains the flexibility to take the right step at each point in a patients journey without worrying about volume. We contractually align incentives with our selected network. Our systems flag over-testing and over-procedure. They compress admission duration. Patients have intelligence-supported and known advocates who close the loop on discharge, adherence, and chronic follow-up. This is where trust is repaid.
4) Data Architecture and Biobanking Layer: Privacy is enforced in infrastructure rather than in policy. The architecture consists of three layers across three legal entities, with three boundaries. Every cross-boundary event is logged and audited. Our approach to biological data is built with retrospective assay as a first principle; we collect biosamples from every member at enrollment. Assays run only on the subset triggered by clinical or research relevance. This is how we build a longitudinal biology database economically.
5) Medical and Scientific Asset Layer: Separate Data Trust which holds de-identified data and biobank custody and signs pharma and academic partnerships without contaminating the insurer’s or the care entity’s profit and loss. It is also the channel through which the Data Dividend flows back to members.
Model Commercialization
Navigation-linked margin capture. The largest and most opaque block of private insurance claims in India is secondary and tertiary care. A trusted navigator reduces unnecessary procedures, routes to better-performing hospitals, negotiates pricing, compresses admissions, and closes the discharge loop. On a hundred-thousand-member book, partial conversion of these levers is worth several hundred crores a year in reduced claims. The revenue can be booked two ways: directly from the member as a subscription that bundles the navigator, and from the insurer as a per-member-per-month management fee or a share of the claims savings demonstrably produced. Both forms activate the same underlying mechanism.
Research partnership revenue. The Data Trust holds four assets with distinct revenue shapes: a metabolomics reference panel, a proteomics and pQTL panel, a CGM glycaemic reference dataset, and — over a longer horizon — a pre-disease trajectory biomarker. Comparable deals establish that the category is real and the magnitudes are material: pharma
partners in the UK Biobank proteomics consortium at 40 - 120Cr each, deCODE’s Roche and Amgen arrangements in the 1000 - 3000 Cr, 23andMe’s GSK partnership at the 2500 Cr mark, and biomarker-licensing valuations in the 10-50 Cr range. The ninety-day sprint will size each asset bottom-up against pharma willingness-to-pay and return with defensible ranges. The Year One milestone is LOIs signed in at least two of these shapes; the Year Four milestone is conversion to revenue. Neither is speculative — both are anchored in deal shapes the global biobank category has already demonstrated.
Proof Points, Phasing and Ask
Minimum Viable Proof: A single cohort of members, run on the full architecture, measured against five legs simultaneously. First, members enroll and engage in the healthy/non-clinical state; trust forms. Second, when clinical events occur, members route through the Jio Arogya navigator rather than around us; trust is operational. Third, deep phenotype produces clinically meaningful risk stratification that changes the timing or the choice of intervention. Fourth, outcomes translate to measurable claims-cost reduction priced by an insurer, whether Jio Allianz when it is live or a group insurer before then. Fifth, de-identified data finds buyers; at least two pharma LOIs are signed inside the first hundred and eighty days. All five in one pilot, not in series. Any one leg failing tells us which piece is weak before we scale.
First customer (deliberately open): Three beachheads are credible and each produces a different economic and data shape. Urban affluent, forty-five to sixty-five, cardiometabolic: a clinical-grade product at three to four lakhs a year, where trust forms through preventive surveillance and navigation monetizes fast because secondary events are near. Urban affluent, thirty to forty-five, performance and longevity: a consumer-grade product at seventy-five thousand to one and a half lakhs a year, where trust forms through daily insights and ambient sensing, and the compounding relationship is the longest of the three. Employer groups, twenty- five to fifty-five: a capitated group policy, where the employer buys trust on behalf of the employee and the scale ramp is the fastest. The same Personal Agent, the same navigator, the same biobank, the same Data Trust, and the same Data Dividend underlie all three. Pre- deciding in April is a bet that should be a decision in July, with the economics and the signed LOIs in hand. The sprint’s output is a chosen beachhead priced and ready to launch, not three candidates and a recommendation.
Build vs partner options: Four ways to assemble the operating system exist, and the decision between them is the most material one we face after the framing itself.1) Asset-light platform owns data, agent, navigator, and Data Trust and partners everything clinical. 2) Asset-medium build owns the data spine and a thin physical footprint and partners labs, primary care, and hospitals. 3) Asset-heavy build owns the delivery chain end-to-end, at the capex and complexity implied. 4) A platform-as-a-service exposes the operating system to third-party insurers, employers, and hospital chains with the thinnest direct consumer presence. Each of the internal
plans in circulation maps onto one of these options. Our strong bias is towards the asset-light or asset-medium platform, but we will validate in the first 90 days.
Phasing Gates:
Gate 1 (end July 2026): Pressure Tested and Fundable Operating Plan
A unified specification across all five operating system blocks built in integration with all internal stakeholders including Jio Allianz. Bottoms up economics and financials with strong conviction. The Data Trust legal structure filed. Two to three signed pharma LOIs. A chosen beachhead with pilot design. A chosen build option. Key early team hires. Early Clinical AI and Personal Agent Prototype.
Gate 2 (end of October 2026): Live Consumer Pilot in the Chosen Beachhead
Five hundred to one thousand members on the full architecture, including the navigation layer. Six months of engagement, navigation, outcomes, and biobank seed in hand by March 2027.
Gate 3 (mid-2027): Integrated Jio Allianz-product and cohort launches
Including Jamnagar with 10,000 families. The scale decision is taken with twelve months of outcomes and signed pharma partnerships already on record.
Ask today: 1) Endorsement of the approach (proposed operating model, asset light leaning, data and clinical AI as core, not after thought) 2) Identified owner for the operating system, reporting into your office. 3) Authorized ninety-day sprint at a small funded scope.