The grievance desk should be able to overrule the claims desk

Every Indian general insurer has a grievance redressal officer, a published process, and timelines set by the regulator for acknowledging and resolving complaints. Most also have a board committee that reviews how policyholders are treated. On paper the machinery is complete. In practice, at many insurers and at Nilakhya until 2022, a complaint about a claim travels back to the claims department, where it is reviewed by people who report to the manager whose team made the decision. The complaint is acknowledged on time, answered on time and, in most cases, rejected on time. The customer who disagrees is told the same thing more politely.

I ran a claims department for five years under that arrangement and did not think it was wrong. I now believe it is the most common reason that complaints which should be settled inside an insurer end up before the Insurance Ombudsman. My argument is simple. A grievance function should have the authority to revise a claim decision, and the people exercising that authority should have had no part in making it. Without that authority, a grievance team can measure dissatisfaction but cannot correct it.

Who answers the complaint

Consider what a complaint about a health claim usually concerns. It is rarely a full rejection; those are fewer, and they are usually argued on the policy's terms. It is more often a partial settlement. The hospital bill was ₹2.4 lakh and the insurer paid ₹1.7 lakh. The difference is made up of items the policy treats as non-payable, consumables billed separately, and a proportionate deduction applied across the whole bill because the patient chose a room above the policy's limit. Each deduction may be defensible under the wording. Together they leave a family with a bill it had not expected and a letter it cannot follow.

When that family complains, the reviewer in the claims department checks whether each deduction matches the wording and the department's practice. It almost always does, because the practice was the department's own. The review is honest. It is also circular. It asks whether the decision followed the rules the decision-makers set, not whether those rules produce outcomes the buyer of the policy could reasonably have predicted.

In 2021–22 nearly two-thirds of the Ombudsman's awards against Nilakhya concerned partial health settlements of exactly this kind. The Ombudsman was not applying a different law. It was asking the question our own review had not asked: whether the customer had been told, clearly and in advance, what the policy would not pay for.

What a revision rate tells a board

In 2022 we moved the grievance team out of claims. It now reports to me, with a line to the chair of our board's policyholder protection committee, and complaints about claims go to a review panel of senior doctors and claims managers who had no part in the original decision. The panel can revise the decision. We also required that any deduction above ₹10,000 cite the clause behind it, in plain words, in the settlement letter.

The most useful figure this produced is the revision rate: the share of reviewed complaints in which the panel changed the outcome. In the first year it was 19 per cent, and the revisions added ₹17 crore to claims paid. By 2025–26 it had fallen to 8 per cent. That fall is the point. It did not happen because the panel became stricter. It happened because every revision was returned to the assessor who made the original decision, with the panel's reasoning, and the claims team began to change its practice before complaints arrived. Over the same period grievances fell from 31 to 14 per 10,000 policies, and the Ombudsman's awards against us from 212 a year to 74.

For a board, the revision rate does what a count of grievances cannot. A count rises and falls with volume, with the season and with how easy the insurer makes it to complain. A revision rate measures how often the company's first answer was wrong in the view of its own senior people. A rising rate in one product or region is an early sign of trouble in claims practice, months before it shows in Ombudsman awards or a regulator's inspection. A rate near zero should worry directors too. It usually means the reviewers are not independent, not that the decisions are perfect.

I would suggest that every insurer's policyholder protection committee see four figures each quarter: grievances per 10,000 policies by product; the revision rate on reviewed complaints; the share of customers who contacted the insurer a second time about the same matter; and the Ombudsman's awards against the insurer, with the reasons grouped. None of them requires a new system. All four require that someone other than the claims department be responsible for them.

The objections worth taking seriously

The strongest objection came from our own chief claims officer, who had succeeded me in that role. A claims department, she argued, must apply the policy consistently. If a panel elsewhere can overturn its decisions, assessors will stop applying the wording firmly and will pay doubtful claims to avoid being overruled. Claims cost will rise, and the honest majority of policyholders, whose premiums fund every claim, will pay for it.

She was right that claims cost rose. The first year cost ₹17 crore in revised payments, against our CFO's estimate of ₹22 crore. She was also right that relations between the teams were poor for several months, and I made that worse by returning the panel's reasons to assessors only from the eighth month. Where I think she was wrong is in assuming that firmness and fairness pull in opposite directions. The panel upheld 81 per cent of the decisions it reviewed in its first year. What it changed were mostly deductions that the wording permitted but that no customer could have predicted from the policy schedule or the sales conversation. Those are not cases of firmness. They are cases in which an insurer relied on fine print it had never explained.

A second objection is that this kind of independence belongs to the Ombudsman, and that an insurer which builds its own appeal layer is duplicating a public institution. I disagree. The Ombudsman exists for disputes an insurer could not resolve. It was never meant to be the place where an insurer discovers that its own practice has drifted. Every complaint that reaches it has already cost the customer months and the insurer its credibility, whatever the award.

A third objection is the cost in people. Our panel takes about eleven senior staff, which is not trivial. I would set that against the alternative: the hours spent defending decisions before the Ombudsman, the renewals lost from customers who felt cheated, and the risk that a regulator's inspection finds the pattern before the board does.

None of this needs a change in regulation. It needs a management decision that the people who answer complaints may change outcomes, and a board that asks to see how often they do. A complaint is the cheapest audit of claims practice an insurer will ever receive. It is wasted if it is reviewed only by the people it is about.