A filled permission field can conceal an unresolved decision

Every required field could be populated, but twenty-eight entries still carried overlapping claims. That was the uncomfortable finding in our review of 120 recording-territory-period entries for a proposed client service. The schema could accept the data more easily than the business could explain what it was entitled to represent as supported. I have become cautious about completion measures since then. A record can be syntactically finished while the decision a customer will infer from it remains unresolved.

Completeness can hide incompatible accounts

Ninety-two entries had a documented scope supported by the responsible sources. The other twenty-eight did not lack all evidence. They contained evidence that required clarification because the claims overlapped. Calling them incomplete would have been understandable in a workflow sense, but it risked suggesting that an operator merely needed to fill a missing value. The difficulty was not an empty box. It was determining what the service could responsibly say while the relevant parties had not supplied a settled instruction. More data of the same ambiguous kind would not automatically resolve that problem.

The account sponsor suggested displaying both claims and allowing the user to decide. That proposal had a commercial rationale. It preserved a broader apparent offering, exposed some of the uncertainty and avoided making our service wait for parties outside our control. In a product designed explicitly to present competing claims for expert investigation, such an approach might have a different meaning. Our proposed supported set, however, would be read as having a resolved scope. Listing the conflict inside that set would transfer a decision to the customer while retaining a presentation that implied we had already dealt with it.

Devika and I therefore supported a bounded ₹58 lakh evidence review. Authorised client representatives and the relevant advisers retained the decisions that belonged to them. Our data team organised the source material and its scope, made the overlaps visible and recorded the instructions supported through that process. We did not decide a legal owner by choosing a preferred record, nor did we alter the underlying agreements. The distinction between facilitating a decision and possessing authority to make it was central to the design of the review.

Nineteen questioned entries received clarification. Nine remained outside the supported set, leaving 111 entries with documented scope for the service under examination. The nine did not disappear from our evidence record. Nor did we assert that they could never be used or that every claim associated with them was invalid. We had a narrower conclusion: the proposed handling was not supported well enough for inclusion in this set at that time. That limitation was less satisfying than a universal answer, but it was the answer our role and evidence allowed.

Keep the uncertainty attached to the offering

The commercial account had to retain both the narrower set and the cost. The ₹58 lakh expense was not transformed into a recovered-royalty figure or a quantified litigation saving. We had not established either. It was the cost of bringing the proposed service's account into line with the evidence available to it. The remaining exceptions required continued attention, and the customer could see a smaller supported offering. An executive narrative that hides those consequences under the word quality would make the decision harder for others to assess.

This experience also changed how I discuss ownership of data work. A catalogue team may own the task of assembling evidence without owning the substantive rights decision. A product team may own the display without owning the permission it appears to communicate. A commercial team may own the account relationship without being able to settle an overlap by agreement with its own preferred contact. Describing those responsibilities carefully prevents a familiar problem: every function believes another has approved the meaning carried by the final service, while no authorised person has actually supplied that approval.

There is a danger in the opposite direction as well. The data team should not become so concerned about its limits that it simply forwards an unstructured conflict and calls its work finished. We can make the issue intelligible, identify the scope that needs clarification, preserve the evidence and explain which instruction is still missing. That is useful professional work. The boundary concerns authority over the substantive decision, not a refusal to help the people who have it. Our review was valuable because it supported their judgement without pretending to replace it.

An exception also needs a route back into the service when its condition changes. Keeping nine entries outside the supported set was not a permanent verdict. Further evidence or authorised clarification could change their status. The record should preserve why an entry was held and what kind of resolution would permit reconsideration, so a later colleague does not either bypass the hold casually or treat it as an unexplained prohibition. The aim is a living account of what is supported, with the uncertainty visible and revisable through the proper process.

Ask what the customer will infer

A director reviewing this kind of issue should begin with the service promise. What will a customer reasonably infer from an entry's inclusion? If inclusion suggests a supported scope, then the board needs to understand how management handles competing evidence before the entry appears there. The question is broader than whether the database contains a warning field. A warning can exist without being carried into the product or the commercial explanation. The board should hear how the enterprise's representation remains coherent across those different encounters.

I would not ask a board to choose between the underlying claims. That would confuse oversight with a specialist decision and may place confidential or legally sensitive material in an inappropriate discussion. I would ask directors to examine the mandate, authority and escalation arrangements by which management obtains a supported instruction. They should also understand the financial and service consequence of leaving the issue unresolved. That connects governance to the actual business choice without turning a committee meeting into a catalogue dispute-resolution desk.

The enduring lesson for me is that evidence does not become permission merely because a system can store it. A service may need to preserve several accounts while declining to present any of them as settled. That can be commercially inconvenient and operationally demanding, but it is more defensible than allowing a completed field to carry an authority the enterprise does not possess. In our review, 111 entries could be supported and nine could not yet be included. Keeping that distinction explicit was the contribution the data function was qualified to make.