Embedded Insurance Commission-Sharing Rules in India 2026: Who May Lawfully Be Paid, and How

Embedded insurance in India has moved well past travel add-ons. Device protection sold with smartphones, transit cover embedded in logistics platforms, health riders attached to lending products, cyber cover bundled into SaaS subscriptions for SMEs, and appliance protection at retail checkout are all live programs in 2026. Behind each one sits the same legal question, and it is not a technology question: who in the value chain may lawfully be paid from the insurance transaction, and for what?

The question matters because the partner that owns the customer relationship (the marketplace, lender, OEM, or platform) usually holds no insurance licence, while the law reserves insurance remuneration for licensed entities. The commercial pressure runs the other way: the partner controls distribution and expects to monetise it. Most compliance failures in embedded distribution are attempts to square that circle with creative payment labels.

The stakes are rising in 2026 for three reasons. First, the Sabka Bima Sabki Raksha (Amendment of Insurance Laws) Act, 2025, whose intermediary provisions took effect on 5 February 2026, restores IRDAI's explicit statutory power to cap distributor commissions and introduces composite licences and 100 percent FDI in intermediaries, signalling an intent to supervise distribution economics closely rather than loosely. Second, the draft IRDAI (Insurance Intermediaries) (Amendment) Regulations, 2026 (June 2026, still draft) would make intermediary income streams from insurers publicly visible in audited schedules. Third, the commission overhaul reported on 3 July 2026, with a consultation paper expected by end July per Chairperson Ajay Seth, targets exactly the passive, low-effort distribution economics that many embedded programs run on. Brokers who design or intermediate embedded programs need the remuneration rules cold, because they are usually the licensed entity holding the risk when a structure fails.

The architecture of Indian law on insurance remuneration is simple even where the details are dense: insurers may pay commission or remuneration only to persons licensed or registered to solicit and distribute insurance. The Insurance Act, 1938 prohibits paying commission to any person other than a licensed agent or registered intermediary, and the 2025 amendments preserve that architecture while strengthening IRDAI's rule-making power over amounts and manner of payment.

The licensed universe for embedded distribution purposes:

• Insurance brokers, remunerated by insurers under board-approved commission policies per the IRDAI (Payment of Commission) Regulations, 2023, and able to charge client fees for advisory services under the 2018 broker regulations.
• Corporate agents, the natural home for platforms and lenders that want to distribute directly: a registered corporate agent may solicit for a limited panel of insurers per line of business and receive commission lawfully.
• Insurance marketing firms and web aggregators, each with defined solicitation perimeters and remuneration governed by the same board-approved-policy regime.
• Point-of-sales persons (POSPs), engaged through insurers or intermediaries for simple, pre-underwritten products, remunerated through their sponsoring entity.

An unlicensed partner sits outside this universe entirely. It may be paid for genuine non-solicitation services (technology, marketing, data processing, platform integration) at defensible arm's-length value, but it may not receive payment that is in substance consideration for soliciting or procuring insurance business. The line is functional, not cosmetic: what was the payment actually for? A payment calibrated to policies sold is remuneration for solicitation whatever the invoice says.

A large share of embedded distribution runs through group policies: a lender takes a group credit-linked health cover for borrowers, a platform takes a group personal accident policy for gig workers, an OEM takes a group protection policy for device buyers. The group route is attractive because the master policyholder needs no licence to hold the policy. It is also where embedded economics most often go wrong.

Indian group insurance rules rest on a long-standing principle: the group organiser or master policyholder must not profit from arranging cover for its members. The master policyholder may recover from members no more than the premium attributable to them plus, where permitted, defined administrative allowances; it is not a distribution channel entitled to a margin on premium. IRDAI's group insurance framework has restated this principle across product regimes precisely because organisers keep testing it.

The patterns that fail the test recur across programs:

1. Premium marking up: the platform collects INR 799 per member for cover whose actual group premium is INR 450, keeping the spread. This is prohibited profit from group administration, and it also misleads the member about the price of insurance.
2. Volume-linked administration fees: an administration charge from the insurer that scales with member count or premium volume rather than with administrative work performed functions as commission to an unlicensed entity.
3. Compulsory bundling with opaque pricing: members auto-enrolled with the insurance cost buried in a composite subscription price, removing any visibility into what was paid for cover.

The compliant structure is unglamorous: the member pays the actual premium (or the organiser bears it), administrative recoveries reflect documented administrative cost, and any distribution economics flow to a licensed intermediary that actually performs solicitation and servicing. Brokers advising group programs should insist on a member-level premium disclosure discipline, because it is the single feature that distinguishes defensible programs in a supervisory review.

Supervisors reviewing embedded programs look past labels to payment mechanics. Four structures account for most disguised commission sharing in the Indian market, and each has a tell.

The scaling technology fee. The insurer or broker pays the platform a per-policy or per-member technology fee for API access or integration. Genuine technology pricing reflects usage, capacity, or development cost; a fee that rises linearly with policies sold prices solicitation, not technology. The tell is perfect correlation between the fee and conversion volume.

The inflated marketing services agreement. The platform invoices for advertising, placement, or co-branding at values far above any market benchmark for equivalent media. The tell is valuation: what would this promotion cost if bought from an unrelated party? The excess over that number is distribution remuneration in disguise.

The data and analytics fee. Payments for customer data or conversion analytics that in substance compensate access to a captive customer base for solicitation. Aside from the remuneration problem, casual monetisation of customer personal data creates independent exposure under the Digital Personal Data Protection Act, 2023.

The rent-for-shelf arrangement. Fixed periodic payments for exclusive placement of one insurer's product in a checkout flow, calibrated informally to expected volume. Exclusivity payments hover closest to legitimacy, which is why they need the strongest arm's-length documentation.

Consequences run through the Insurance Act, 1938 penalty provisions, where contraventions of the remuneration sections carry monetary penalties that can reach INR 1 crore per violation, alongside directions to unwind arrangements and the supervisory consequences for the licensed entities involved. The rebating prohibition adds a second edge: passing insurance economics to the customer as inducement is separately barred, so a platform cannot cure an unlawful payment stream by handing it to buyers as a discount on the insured product bundle.

Two 2026 developments converge on the same effect: embedded remuneration structures are about to become far more visible.

The draft IRDAI (Insurance Intermediaries) (Amendment) Regulations, 2026, published in June 2026 and still at draft stage, would require brokers, corporate agents, insurance marketing firms, and web aggregators to disclose intermediation revenue and other income received from insurers in a separate schedule to their financial statements, file audited financials with IRDAI by 30 September each year, and publish them on their websites, with stricter disclosure once commission income crosses INR 10 crore. The phrase to sit with is other income from insurers: the schedule is designed to capture precisely the technology fees, marketing recoveries, and service income that embedded structures use alongside headline commission. An intermediary fronting an embedded program would publish, annually and audited, the full shape of its insurer-derived economics.

The commission overhaul reported on 3 July 2026 points the same direction from the conduct side. Effort-based remuneration, paying more for advisory, documentation, and claims servicing than for passive channels, is a direct challenge to embedded economics built on auto-attach flows with negligible service content. Caps by product type, tenure, and complexity would bite hardest on the simple, high-volume covers that dominate embedded distribution. Tighter disclosure is on the same reported list. All of this remains proposal pending the consultation paper expected by end July 2026, but the direction of travel is one-way: payment flows around insurance distribution are being pulled into the light.

For embedded programs the planning assumption should be that by FY2027-28, every material payment stream between insurer, intermediary, and platform is either disclosable, capped, or benchmarked against service effort. Structures that only work in the dark should be treated as already broken.

Compliant embedded distribution is achievable, and the compliant versions are also the durable ones. The design choices reduce to a small set.

License the partner where the partner solicits. If the platform's checkout flow presents, recommends, or defaults customers into cover, the platform is soliciting and should hold a registration: corporate agency is the usual fit, and the composite licences introduced by the 2025 Act make multi-segment registration simpler than before. Registration converts the platform's distribution income from a structuring problem into lawful commission under the insurer's board-approved policy.

Or keep solicitation with the licensed intermediary. Where the partner will not take a licence, the broker or insurer must genuinely own solicitation: product presentation, disclosures, suitability logic, and servicing run under the licensed entity's control, and the partner's role is confined to providing the venue and technology. The partner is then paid for exactly that, at benchmarked arm's-length value, on pricing metrics (integration fees, usage-based platform charges) that do not track policies sold.

Run group programs on cost-true administration. Member-level premium transparency, administrative recoveries tied to documented cost, and distribution economics flowing only to licensed entities.

Document the value of every partner payment. A benchmarking file for each service agreement (what the service is, what it would cost from an unrelated provider, why the price is fair) is inexpensive to build at signing and nearly impossible to reconstruct credibly during an inspection.

For brokers, the 2026 environment is an opening as much as a constraint. Platforms facing disclosure and effort-based scrutiny need exactly what a broker sells: licensed solicitation, documented advice and servicing, claims handling, and clean remuneration plumbing. The broker who can show a platform a fully disclosable program design, and evidence of servicing effort behind every rupee of remuneration, is offering the one embedded architecture that survives every version of the coming rules.