The product stopped selling; the undertaking did not

## A closed book still contains live decisions

The phrase closed product can create a misleading sense of completion. Sales have ended, the current proposition has moved elsewhere and management attention follows the new offer. Yet the contracts already issued may require service for many years. Their payment schedules, benefit events and options remain part of the company's undertaking. When a technology team proposes retiring the system that carries those contracts, the decision concerns more than an old application. It concerns the organisation's ability to understand and perform obligations that were created under an earlier commercial and operating arrangement. That responsibility survives the people who remember the launch.

At the fictional Vaqruni, a legacy-platform retirement proposal initially emphasised transferring records and closing the application. The direction was reasonable. Keeping an ageing system running required specialist support, cost money and retained risks of its own. Operations then identified eighteen older policy variants whose service behaviour needed explicit treatment. A record could be present in the replacement while an important action remained unclear. The problem was not a missing line in a migration total alone. It was whether the receiving arrangement could support the contract when a real servicing event required someone to interpret and act upon it.

I found it useful to separate three questions. Have the relevant records been transferred correctly? Can the replacement perform the required service? Can colleagues explain and operate that service when the event occurs? The questions are connected, but success on the first does not settle the other two. A reconciliation is necessary evidence, not a complete account of continuity. Equally, an experienced colleague's confidence cannot replace a reliable record or an implemented process. The retirement decision needs these forms of evidence brought together by the owners responsible for them, rather than compressed into one technical completion label.

The history matters because product variations can appear minor until a particular circumstance makes them consequential. A payment schedule or benefit treatment may be uncommon and still be central to the person whose contract contains it. The business cannot infer that an infrequent variant is unimportant merely because most records behave differently. Nor should every historical variation automatically demand a separate application forever. Management must identify what the contract requires, how the replacement will support it and where professional interpretation is needed. That is more precise work than preserving the old system simply because it is familiar.

## Test the service, not only the transfer

Our review used ninety representative service scenarios. At the first decision point, seventy-three had been demonstrated on the replacement and seventeen needed further work. We kept those as counts of scenarios. It would have been misleading to describe the result as a percentage of policyholders protected or a percentage of contractual risk removed. The scenarios differed in frequency and consequence. One unresolved treatment might affect relatively few contracts but matter greatly when a benefit became due. A board needs enough explanation to understand that distribution of significance, rather than a neat progress percentage detached from the work being assessed.

The demonstration also needed distinct professional ownership. Operations understood the servicing task, technology the implementation, actuarial colleagues the relevant product and benefit treatment, and legal colleagues the contractual interpretation where required. Bringing those views together did not mean asking everyone to sign the same statement regardless of their responsibility. I wanted each owner to state what had been examined and what remained outside that examination. A collective signature can create false comfort if it hides different assumptions. A clear account of separate responsibilities makes disagreement visible early enough for management to resolve it before the old capability disappears.

There was a reasonable objection from technology: running two arrangements at once adds coordination and support risk. I did not want the review to portray engineers as impatient and operations as uniquely prudent. The proposed retirement addressed a real problem, and continued operation was not a neutral position. The question was whether the evidence available at that point supported retirement, and what bounded additional work would change the answer. That framing helped us avoid a contest between professional groups. It made the chief executive responsible for choosing a course with acknowledged costs, while allowing specialists to retain their assessments.

The board approved ₹2.8 crore of additional legacy operation while the outstanding treatments were resolved. The older platform therefore remained in service, and our account did not claim a completed retirement. That outcome is less satisfying than a finished milestone, but it is more useful to a reader assessing judgement. The company paid for time to complete specific work. It did not purchase the disappearance of all uncertainty. Management still had to demonstrate the remaining treatments, maintain the old arrangement and return for a decision. A delay only helps if the work performed during it addresses the reason the original decision could not yet be supported.

## Give delay a cost and a decision date

There is a temptation to describe caution as though it has no opportunity cost. An executive can postpone a difficult change, retain existing capability and appear to have protected the business. But the expense, support burden and continuing risks do not vanish. I required a renewed review date and a precise account of the unresolved service treatments. The ₹2.8 crore belonged in the decision, alongside the case for not retiring prematurely. It was not a saving, and it was not evidence that every future postponement would be justified. The company needed to reassess the balance as the work progressed.

This is where board oversight can be especially useful. Directors can ask whether management has defined the remaining decision, whether the professional owners agree on its scope and whether the proposed time is credible. They should also ask what would happen if the remaining work took longer. That enquiry need not turn the board into a technical steering group. It can remain focused on the undertaking, the evidence and the consequences of the available choices. A clear executive account gives directors room to challenge without requiring them to inspect every application setting or resolve specialist disagreements beyond their competence.

The lesson extends beyond insurance, although the contractual details differ by sector. A company may stop selling a product, replace a service or leave a market while responsibilities continue. Records, support and expertise have to survive long enough to honour those responsibilities. The correct answer is not always to retain the old platform. It may be a well-supported replacement, a specialist servicing arrangement or another lawful and appropriate route. The essential discipline is to identify the continuing undertaking before selecting the technical endpoint. Otherwise the organisation risks defining completion around the asset it wishes to remove rather than the people it still has to serve.

I want a retirement paper to tell me what becomes possible after the change and what must remain possible despite it. That pair of questions keeps the benefit of modernisation in view while preserving continuity of obligation. At Vaqruni, it led to additional cost and an unfinished transition account with specific work still due. Another programme may have enough evidence to proceed immediately. Both decisions can be responsible. What would be difficult to defend is treating a transferred record as proof that every future service has been understood, or allowing the end of new sales to stand in for the end of the company's promise.