The Registry Would Publish Your Conduct Record: Reading IRDAI's 1 September PIR Paper as a Broker

IRDAI issued a consultation paper on 1 September 2026 proposing a Public Insurance Registry as digital public infrastructure for insurance, spanning insurers, intermediaries, reinsurers, financial institutions, government agencies and policyholders. The registry has been discussed in the industry for most of this year, and the earlier framing was largely a records story: one authoritative, consent-based ledger of policies, claims and grievances, sitting alongside the transaction layer, ending the reconciliation drudgery that eats broker back offices.

This paper adds something the earlier discussion did not put in writing. It proposes a customer-facing intermediary layer. The paper says the PIR could enable customers to know their insurance intermediary or agent, and to access information on sales quality, persistency, complaints, mis-selling, disciplinary action and blacklisting.

Read that list slowly. Four of those six items are conduct history, and two of them (disciplinary action and blacklisting) are outcomes a firm cannot restate or contextualise once they are on a screen. The registry stops being a plumbing upgrade at that point and becomes a distribution variable.

The strategic sentence in the paper is separate from the intermediary list and worth quoting in full. IRDAI wants the PIR to enable product comparison, suitability assessment and a digital purchase journey, shifting customers towards informed, low-cost purchases rather than intermediary-assisted, high-cost sales. That is a statement of intent about distribution economics, published by the regulator, in a paper about data infrastructure.

The design described in the paper is federated rather than centralised. Records stay with the institution that holds them, under source-system primacy, and the registry resolves and links rather than warehousing a second copy. Access is purpose-linked and role-based, and consent is described as specific, informed, revocable and auditable.

That architecture has three practical consequences for a broking firm.

1. You remain the source system for what you originate. Federation means the registry points at your records, so the quality of what a customer or a regulator sees is the quality of what sits in your own database on the day they look. There is no central cleanup team between your CRM and the screen.
2. Migration is event-driven, not a bulk load. The paper describes existing records being migrated progressively at renewal or other material events. Your book therefore enters the registry one renewal at a time, over a full policy cycle, and the state of a record at its next renewal is the state that gets published.
3. Linkage is the hard part. The registry would link product details, policy and claims records, intermediary information, grievances and policyholder identity. Every one of those joins depends on identifiers matching across systems that were never built to agree with each other.

Event-driven migration is the detail most firms will misjudge. It sounds gentle, and it removes the comfort of a single cutover date to prepare against. If your April renewals migrate in April, the hygiene work has to be finished before your busiest month.

IRDAI has positioned the PIR as furthering the objectives of the Sabka Bima Sabki Raksha (Amendment of Insurance Laws) Act, 2025, which is the same statute reshaping intermediary licensing and structure. The registry and the licensing regime are being built to fit each other, which is a reason to treat the paper as a direction of travel rather than a trial balloon. Our note on what the Sabka Bima Act changed for intermediaries covers the licensing side.

The paper names six categories of intermediary information. None of them is defined in the paper, which is precisely why the comment window matters. Here is the honest reading of what each one would draw on and where a commercial broking firm is exposed.

Sales quality. The least defined and the most consequential. In a life context it usually means some blend of early lapse, free-look cancellation and complaint incidence per thousand policies. In commercial lines there is no settled measure, so whatever gets defined will be defined without commercial broking in the room unless commercial brokers write in.

Persistency. A renewal-retention ratio measured at fixed intervals. Commercial brokers who place annual policies through multiple insurers will find their persistency reads as churn whenever a client moves insurer on the broker's own advice. A client remarketed from one insurer to another at better terms is good broking and, on a naive persistency measure, a non-persisting policy.

Complaints. Volume, and almost certainly ratio to policies serviced, drawn from the grievance ledger. This is the metric most within your control and the one where firms are weakest, because complaint records tend to live in email rather than in a system.

Mis-selling. Complaints closed with a mis-selling classification, or upheld against the intermediary. The classification is applied at closure, often by an insurer's grievance team, and often without the broker seeing the category that was assigned.

Disciplinary action. Regulatory orders against the entity or its principal officer.

Blacklisting. Insurer-level or regulator-level exclusion.

Commercial insurance is not bought off a comparison screen, and the PIR does not change that. A large property programme still needs structuring, wording negotiation, capacity assembly and claims advocacy. What changes is the shape of the conversation that decides who gets the mandate.

Today a corporate buyer choosing between three brokers relies on the pitch, on references, and on whatever procurement can verify. With a registry screen, the same buyer opens a public record before the pitch.

• Procurement screening. A public conduct record is machine-readable evidence, so it enters a procurement scoring matrix as a criterion. Firms with a clean record will attach the screen to the bid. Firms without one will be asked why they did not.
• Renewal challenge. An incumbent broker with a rising complaint ratio gives the buyer's finance team a documented reason to run a tender, and a documented answer if anyone later asks why the tender was run.
• Panel review. Corporates that maintain a broker panel will re-run the screen at each panel review, which converts conduct into a recurring test rather than a one-time check.
• Board and audit questions. For a listed buyer, an intermediary with a disciplinary entry is a documented third-party risk the audit committee can be asked about.

The firms that gain from this are the ones whose service quality has always been better than they could prove. That has been the structural problem in Indian commercial broking: quality is invisible at the point of selection, so price competes and service does not. A public record makes service legible, and legibility rewards the firms already doing the work. The exposure runs the other way for firms carrying an unmanaged grievance backlog, because a complaint that sat open for 90 days while it moved between an insurer and a servicing team looks identical on a registry screen to a complaint that was ignored.

Records migrate at renewal, so a firm's numbers on day one are the numbers it has been accumulating for the twelve months before its renewals migrate. Work backwards from that.

Quarter one: instrument what you cannot currently measure

Move every grievance out of email into a single register with a unique reference, an intake timestamp, a closure timestamp, an assigned category and a named owner. Most mid-sized broking firms cannot today produce a complaint ratio on demand, which means they cannot see the number the registry would publish. Reconcile that register monthly against each insurer's grievance data for your code, because the insurer's count is the one likelier to feed the registry, and the two counts will not match at the first attempt.

Do the same for persistency. Produce a renewal-retention number by client, by line and by insurer, separating three cases a naive ratio conflates: the client who left, the client who stayed but changed insurer, and the policy consolidated into another programme.

Quarter two: close the backlog and fix the categories

Work the open grievance backlog to closure with documented outcomes. For every complaint closed in the last twenty-four months, obtain the closure category the insurer applied and challenge the ones that are wrong now, while the record is still administrative rather than public. This is unglamorous and it is the highest-return work in the programme.

Quarters three and four: evidence and identifiers. Build the servicing evidence trail: dated proof of quote comparison, of suitability discussion, of policy delivery, of endorsement turnaround, of claim intimation and follow-up. Sales quality, whatever definition lands, will be assessed against records like these. Our broker data-readiness note sets out the record-level detail.

Then fix identifiers. In a federated registry, a record that does not join is a record that does not exist, or worse, joins to the wrong entity.

The registry links product, policy, claims, intermediary and identity records. In commercial lines, the entity you are joining on is a company, and Indian company data is messy in specific, predictable ways.

Run these five checks across the book before the next renewal cycle.

1. GSTIN validity and state. A group with operations in eight states has eight GSTINs. Policies issued against a head-office GSTIN for risks located elsewhere will join to the wrong registered entity. Capture the GSTIN that matches the risk location, and store it as a validated field rather than free text.
2. CIN against the legal name. Match every corporate client's CIN to the exact name on the MCA record, not the trading name on the policy schedule. Subsidiaries with near-identical names are where the joins fail silently.
3. Group structure. Record parent and subsidiary relationships explicitly. A claim on a subsidiary that shows against the parent, or the reverse, distorts both entities' records.
4. Your own identifiers. Your broker registration number, PAN, GSTIN and principal officer details should read identically in every insurer's system. Variations create a fragmented conduct record where your numbers are split across near-duplicate versions of your firm.
5. Policy and endorsement keys. Every endorsement must carry the parent policy reference. Endorsement drift is the most common cause of premium and sum-insured mismatches between broker and insurer books.

Consent that is specific, informed, revocable and auditable

That phrasing from the proposed architecture has an operational consequence brokers should plan for now. Consent must be captured per purpose and be withdrawable, which means a single blanket clause in a client mandate letter will not carry. Rewrite the client mandate to record registry-related consents separately and to log each grant and withdrawal with a timestamp. A certificate of insurance or a policy schedule proves cover; it does not prove consent, and the registry would ask for both.

The paper seeks comments on objectives, data architecture, identity framework, privacy safeguards, governance and implementation readiness. Commercial broking is the constituency least likely to file, because the intermediary discussion in India defaults to individual agents and life distribution. Four points are worth putting on the record.

Define the metrics separately for commercial lines. Persistency, sales quality and mis-selling all carry retail-life assumptions. Ask for either a commercial-lines definition or explicit exclusion of commercial placement from metrics designed for retail suitability. Give the remarketing example: a client moved to a better insurer on advice should not read as a persistency failure.

Ask for an intermediary right of response. Mis-selling and complaint categories are assigned by others and published against you. Ask for notification when a category is assigned, a defined window to contest it, and a visible annotation when a category is disputed or overturned.

Ask for an ageing rule on conduct data. A complaint from 2021 and a complaint from this quarter should not carry the same weight on a public screen. Ask that displayed metrics use a rolling window and that resolved and superseded entries age off.

Ask for pre-publication access to your own record. Before anything is customer-facing, an intermediary should see exactly what would be shown, so errors are corrected against the source system rather than argued about in public.

File as a firm, and file specifics. A letter arguing that the registry threatens intermediaries will be discounted, because the paper has already stated its distribution intent. A letter proposing a workable persistency definition for annual commercial placements is a drafting contribution, and drafting contributions are what consultation papers absorb.

The distribution intent in the paper is not a reason to oppose the registry. Commercial broking earns its fee on structuring, wording and claims outcomes, which is exactly the work a comparison screen cannot do. The firms with something to fear from a published conduct record are the ones whose fee rests on the customer not being able to check. Our wider PIR briefing for commercial brokers sets out the registry's full scope, and the intermediary disclosure pack note covers the parallel disclosure workstream.