Thirty-seven insured sites looked like thirty-seven separate entries in the account. Each had an address, an operating description and its own relationship with the insurer. In our 2022 industrial-park review, twenty-four of those sites relied on the same two access roads and one power substation. The addresses remained different, but part of the operating dependency was shared. The finding did not tell us what loss would occur in a future event. It told us that a location-by-location description was leaving out a connection relevant to understanding the obligations we had accepted.
The combined sum insured across all thirty-seven sites was ₹486 crore. That figure needed careful handling. It was not an estimate of probable loss, the amount a reinsurer would pay or the capital required to support the business. Presenting it next to a map could make those interpretations feel plausible even without anyone explicitly stating them. I wanted the board discussion to retain the meaning of the amount and the limits of the exercise. An important concentration question should not become more persuasive by quietly changing the meaning of the numbers used to describe it.
We allocated ₹86 lakh to mapping and engineering review. Risk head Anjali Qevzemi connected the enterprise question with the relevant specialists. Underwriting, actuarial, engineering and reinsurance colleagues retained their assessments and decision rights. My task was to ensure that the shared dependency received resources and attention across organisational boundaries. It was not to prescribe terms from the map or to assume that every operation would be affected in the same way. The connection justified further work; it did not supply the answers that work was meant to develop.
## Counting locations does not describe their connections
Separate summaries are useful. They allow an organisation to understand a site in enough detail to make decisions about it, assign responsibility and maintain a relationship. The difficulty arises when those summaries are added together as though the separation of records establishes the independence of what they describe. A physical network does not follow the boundaries of an administrative record. Roads, power, water, communications and suppliers can connect operations whose individual accounts are otherwise complete. The issue is not that the records are necessarily wrong. It is that a different question needs a view across them.
The two roads and one substation in this fictional case were concrete dependencies, not a general statement that every industrial park is equally exposed. Individual sites could have different arrangements, alternative access, operating tolerances or consequences from interruption. Those distinctions mattered. A shared connection can be a reason to investigate without becoming a reason to declare identical vulnerability. The enterprise account needed both ideas at once: the sites were connected, and the practical significance of that connection still required assessment. Losing either idea would make the presentation simpler at the cost of making the judgement less useful.
I also wanted the review to avoid turning an absence of visible connection into proof of independence. We had identified a particular relationship among twenty-four sites. That did not establish that the remaining thirteen had no common dependencies of another kind. Nor did it mean that the thirty-seven-site exercise described every concentration in the insurer. Scope was part of the finding. Directors should be able to understand what management examined and what remains outside the account, rather than infer a complete map of the enterprise from one well-presented example.
## A map creates questions for the right disciplines
A visual representation can be powerful because it makes a relationship easier to see. That strength can also encourage people to believe they understand more than the evidence supports. Seeing several sites linked to the same infrastructure does not determine interruption frequency, duration, policy response or financial severity. Each of those questions involves different facts and professional work. I wanted the map to connect those disciplines, not replace them. The chief executive can insist that a cross-cutting question has an owner and resources while leaving the technical assessment with people authorised and equipped to make it.
The ₹86 lakh programme therefore belonged in the account as investigation expense. It was not a premium saving or a quantified avoided loss. We did not invent a catastrophe that failed to occur in order to claim that the review had paid for itself. The immediate result was a clearer question and a funded route for examining it. That can be a legitimate executive outcome. It changes the basis on which later decisions will be made, even though it does not provide a neat financial return at the moment the programme is approved.
There was a practical coordination cost as well. Colleagues had to reconcile site accounts, understand the relevant infrastructure and make their different assessments usable to one another. Work that is obvious in hindsight can still compete with other responsibilities when it is first proposed. The executive decision needed to recognise that competition rather than treat a request for better understanding as costless. Funding the work made the question consequential. It did not remove the need to decide what evidence was sufficient for the next step or how unresolved uncertainty should be carried into that decision.
## Preserve the uncertainty when presenting the concentration
A board paper can lose useful precision when it is shortened. The thirty-seven sites become a portfolio, the twenty-four connected sites become a concentration and ₹486 crore becomes an exposure number without its original meaning. Each abbreviation may appear harmless. Together they can create an impression that the loss has already been calculated. I would rather retain an extra sentence explaining the population and measure than gain a more dramatic headline. A director should not have to reconstruct the distinction between sum insured and a scenario assessment from a footnote after the discussion has moved on.
The same care applies to the next action. Management can report that mapping and engineering review are funded, underway or complete only as the evidence supports each state. A completed map does not mean that every resulting decision has been made. A revised technical assessment does not imply that all affected relationships have changed. Keeping those stages separate helps the board understand what remains to be resolved and by whom. It also prevents the executive sponsor from claiming the eventual outcome of work that other professionals have yet to perform.
My contribution in this case was to make the connection an enterprise responsibility. I had learned through regional and operating roles that the boundary of a reporting unit can conceal work shared with another unit. The industrial-park review gave that lesson a physical form. The sites were distinct, the roads and power dependency were shared, and the consequence required qualified assessment. Holding those facts together was more useful than choosing between a reassuring diversification story and an exaggerated concentration story.
A reader considering their own enterprise can begin with a modest question: what relationship disappears when our separate accounts are added together? The answer may concern infrastructure, people, suppliers or another dependency entirely. It should be investigated on its own evidence rather than forced into this fictional example. For an insurer, the question belongs alongside disciplined technical work and a clear account of obligations. For a chief executive, it is a reminder that the organisation's reporting structure is a way of seeing the business, not the boundary of the business itself.