Ask a building when its rooftop unit was last serviced and the honest answer is usually a search. Somebody looks through an inbox, somebody else checks a filing cabinet, a third person phones the contractor they think did it, and the person who would have known left eighteen months ago.
The fix is not a better filing system, because filing systems describe things rather than being attached to them. It is putting the record on the thing itself: the unit carries its own install date, its warranty term, its service history and what it has cost the building so far.
That changes retrieval completely. Finding the record means pointing at the appliance — in the twin, or by scanning the tag physically stuck to it — instead of reconstructing from memory which document referred to which piece of equipment on which roof.
It also changes what survives. Staff turnover, a change of management company, even the sale of the building: none of them take the history away, because the history was never sitting in any individual's inbox or on any one laptop to begin with.
The value shows up first in arguments. Peeling paint at an entryway two years in is an entirely different conversation when the surface itself names the contractor who applied it, the month it went on, and the two-year workmanship warranty that expired four months ago.
It shows up again in spend. Total cost of ownership per asset is only a meaningful number when every repair was logged against that asset rather than against a cost centre, and it turns into a real figure instead of an allocation somebody made at year end.
And it shows up at replacement, which is where the money is. An asset that knows its own age and its own repair record can tell you whether the next failure is worth fixing or whether the unit has reached the point where repairing it is simply funding the next failure.
Patterns emerge that are invisible in a work order list. The same unit patched three times in two years does not look like anything in a queue sorted by date; against the asset it is an obvious case for replacement, and usually a cheaper one than the fourth patch.
None of this requires the building to have a model. The record hangs off the asset itself, and a twin, where one exists, simply gives you a second and faster way to reach the same record — by clicking the thing in three dimensions instead of searching for its name.
The discipline required is genuinely small, which is why it is worth insisting on: log the work against the asset, every time, including the small jobs. The small jobs are the ones that later turn out to have been the early warning.
What accumulates from that discipline is the most valuable thing a property owns and the thing almost nobody keeps — an account of what this specific building has actually done, rather than what buildings of its type are said to do.