The placement desk is being rebuilt while the cession rules stay put
Reinsurance operations in most Indian general insurers still run on a strange split. The capacity panel is getting wider and more international by the quarter, yet the placement workflow that connects a primary insurer to that panel is often a chain of emails, attached spreadsheets, and PDF slips moving between an underwriting desk, a reinsurance broker, and a treaty section. The contrast is now hard to ignore.
Two developments make 2026 the year this gap starts to close. First, GIC Re has begun pairing with digital reinsurance providers, and the global placement platforms that brokers already use abroad, such as eReinsure and Verisk's Whitespace, support JSON and XML messaging and open APIs rather than re-keyed spreadsheets, so the technology to run placement as structured data is now off the shelf. Second, IFSCA's GIFT City reinsurance hub keeps adding licensed capacity: international insurance offices (IIOs), several of them branches of large global reinsurers, have been licensed in recent quarters, and IIO gross premium has grown sharply year on year, most of it reinsurance.
Wider panel plus more data rails should mean faster, cleaner placement. The catch is that the regulatory spine has not moved in step. The obligatory cession to GIC Re sits at 4% for FY2026-27 (policies attaching 1 April 2026 to 31 March 2027), GIC Re remains the sole statutory recipient, and the order-of-preference rules still decide who gets shown a risk first. So the broker and the cedent are automating a process whose mandatory first step, and whose lead terms on most domestic treaties, are still set by one counterparty. The interesting work for 2026 is not buying a platform. It is knowing which parts of the placement a rail genuinely improves, which parts the regulation keeps fixed regardless of the technology, and where a faster workflow can quietly create a compliance or coverage problem if nobody has thought about the sequence. Understanding what the rails change and what they leave untouched is the whole job.
What a digital cession rail actually does to the workflow
Strip away the vendor language and a reinsurance placement platform does four concrete things. It holds the risk submission as structured data rather than a free-text slip, it routes that submission to selected markets with version control, it captures quotes and line sizes back into the same record, and it produces the closing and the bordereaux from data already entered rather than from a fresh spreadsheet.
The operational payoff is in the boring parts. On a facultative placement today, an Indian cedent often sends the same risk detail to GIC Re, to one or two FRB branches, and to a broker who then approaches cross-border reinsurers, each in a slightly different format. Every re-keying is a chance to drop a sum insured digit, mis-state a deductible, or quote an outdated wording reference. A structured submission means the occupancy, the PML basis, the sum insured, the period, and the lead wording travel as fields, not prose, and every market sees the same version.
The single biggest data-hygiene win is not speed, it is that the risk detail a reinsurer prices against and the detail that later appears on the bordereaux and the claim advice are the same record. Mismatches between placement data and bordereaux data are a routine cause of disputed recoveries, and a structured submission closes that gap at source.
For treaty business the gain shifts to the bordereaux itself. Premium and claims bordereaux that today arrive as monthly spreadsheets, each cedent in its own column order, can move to a shared schema (ACORD-style or a platform's PLACEDATA-style template) so the reinsurer ingests them without manual mapping. That matters most for proportional treaties and for any arrangement where profit commission is calculated off the bordereaux, because a clean, machine-readable account closes faster and argues less.
The obligatory cession is the one step you cannot automate away
Every digital placement design in India has to start from a fixed point: 4% of the sum insured on almost every general insurance policy is ceded to GIC Re, by regulation, for FY2026-27. Terrorism premium and amounts ceded to the nuclear pool are carved out at nil. IRDAI has eased the earlier cap on sum insured for the obligatory cession, and the profit-sharing arrangement with GIC Re continues on a broad 50:50 basis with minimum ceding-commission rates set by class, lower for high-loss-ratio lines such as motor third-party and oil and energy and higher for most other classes. The exact figures are reset by IRDAI each year, so the discipline is to read them off the current order rather than carry last year's numbers forward.
What does automation do here? It does not remove the cession, it makes it accurate and timely. The obligatory slice is a deterministic calculation off the policy record, so a well-built rail computes it at booking, posts it to the GIC Re account automatically, and reconciles the commission and the profit-share inputs without a treaty clerk rebuilding a spreadsheet every quarter. The error this prevents is real money: under-ceding or mis-stating commission on the obligatory book is a compliance and reconciliation headache that surfaces at year-end.
The practical instruction for a cedent's reinsurance head is blunt. Map every class of business to its obligatory treatment (including the terrorism and nuclear nil cases) inside the platform on day one. Get that wrong and every downstream bordereau inherits the error.
Order of preference: the sequence the rail must encode
Digital speed is dangerous if it lets a broker show a risk to the wrong market first. India's cession order of preference is not a guideline, it is the compliance frame for who sees a placement and in what sequence, and any automated routing has to encode it.
The broad hierarchy that cedents and brokers work to is: the obligatory cession to GIC Re first, then Indian reinsurers and FRBs (foreign reinsurer branches) for the balance, then GIFT City IIOs, and only then cross-border reinsurers (CBRs) operating purely on a cross-border basis. GIFT City IIOs that retain and reinvest their Indian premium domestically sit ahead of plain CBRs in the queue, which is exactly why the IFSCA hub is pulling capacity onshore. A placement platform that routes a risk to a CBR for terms before the higher-preference markets have been given a fair opportunity does not just breach etiquette, it can leave the cession open to challenge.
The broker-side discipline is to configure the routing logic so that quote requests go out in preference order, with timestamps, and the audit trail shows that domestic and FRB capacity was approached on the same terms before cross-border markets were bound. This is one area where digitisation genuinely helps the regulator's intent rather than working against it. A structured platform records who was shown what, when, and at what price, so a cedent can demonstrate that the order of preference was respected. The old fax-and-email file rarely proved that cleanly.
For reinsurance brokers the read-across is commercial as well as compliant. If your placement system makes the preference sequence visible and auditable, you can defend a placement to the cedent's board and to IRDAI without reconstructing a paper trail after the fact.
Where GIC Re's lead terms still set the market
Automation changes how a risk is shown and accounted for. It does not, on its own, change who sets the price and the wording. On a large share of domestic treaty and facultative business, GIC Re is still the lead, and the lead reinsurer's terms, the rate, the deductible structure, the event limits, and the wording clauses, become the reference the following market accepts or declines.
This is the part cedents sometimes hope technology will dissolve, and it will not. A wider GIFT City panel and faster rails give a cedent more following capacity and arguably better competition for the non-obligatory layers, but on programmes where GIC Re leads, the lead terms anchor the placement. A digital platform that lets you canvass twelve markets in a morning still has to record the lead line and propagate the lead wording to every follower, or you end up with following lines written on subtly different terms, which is a claims dispute waiting to happen.
The useful framing for a reinsurance broker is to separate three things the platform must keep straight: the obligatory cession (fixed, GIC Re, 4%), the lead terms (often GIC Re on domestic business, and binding on followers), and the following capacity (where the wider IFSCA and cross-border panel competes). Each behaves differently and each needs its own field discipline.
The non-negotiable rule sits inside that third category. Following lines must be written on the lead's terms or on a documented difference-in-terms basis. A placement rail that silently lets followers attach their own wording variations is creating latent coverage gaps that only appear at the claim, when the follower argues its narrower wording and the cedent is left carrying the difference. Get this structure right and the technology earns its keep. Get it wrong and a faster placement just produces a faster mess.
Bordereaux data is where the recovery is won or lost
Reinsurance recoveries do not fail at placement, they fail at the bordereau. When a claim hits a proportional treaty, the reinsurer pays against the account it can reconcile, and a bordereau that does not tie back to the original placement data is the single most common reason a recovery stalls.
This is the strongest argument for moving to structured, schema-based bordereaux rather than free-form spreadsheets. If the risk-level data captured at placement (policy reference, period, sum insured, ceded share, premium, lead reference) is the same data that populates the premium and claims bordereaux, reconciliation is a lookup, not an investigation. When the placement record and the bordereau are built separately, in different formats, by different teams, the gaps multiply: a policy on the bordereau that the reinsurer never saw at placement, a claim coded to the wrong treaty year, a ceded share that does not match the slip.
For an Indian cedent the practical steps are concrete. Standardise the bordereau template across all treaties so every reinsurer ingests the same fields. Reconcile the bordereau against the placement record before it goes out, not after a recovery is queried. And keep the claims bordereau coded to the treaty year and the lead reference so a reinsurer can match a loss advice to the exact contract that should respond.
The single control that holds all of this together is a monthly three-way tie-out: policy system to placement record to bordereau. Most ceded-premium leakage and most disputed recoveries are caught in that reconciliation, and a structured rail makes it a report rather than a fortnight of spreadsheet archaeology. The payoff compounds. Clean bordereaux speed up profit-commission settlement, shorten recovery cycles, and make the cedent a more attractive counterparty at renewal.
A practical adoption sequence for cedents and brokers
Most Indian insurers and reinsurance brokers will not replace their reinsurance operations in one go, and they should not try. The sensible path is to digitise the parts that pay back fastest and carry the least placement risk, then move up.
Start with the deterministic, low-judgement work. The obligatory cession calculation, the GIC Re account posting, and the class-wise minimum-commission figures are pure arithmetic off policy data, so automate them first. Next, standardise the bordereau schema across treaties, because that single change improves every recovery and every profit-commission settlement without touching how risks are placed. Only after those foundations are clean should a firm move facultative submissions and treaty quote-routing onto a structured platform, because that is where judgement, market relationships, and the order of preference all interact.
Three questions separate a governed rail from a fast email. Does it generate the obligatory cession and bordereaux straight from the system of record, or does someone still rebuild them in Excel? Does it encode and timestamp the order of preference so a placement is auditable? And does it propagate the lead wording to every following line so followers cannot drift onto their own terms?
For brokers, the commercial case writes itself. A cedent that gets clean bordereaux, defensible preference trails, and faster profit-commission settlement is a cedent that renews. The firms that treat reinsurance operations as a data discipline, not a correspondence habit, will place more business with fewer disputes as the IFSCA panel keeps widening through 2026.