As-Builts and What Happens When Nobody Updates Them

Why this matters

Every building an institution owns has a set of drawings, and in almost every one of them the drawings and the building disagree. The disagreement is not random. It grows, one undocumented change at a time, and the cost of it lands on whoever shows up next holding a flashlight, billing hours to find something that a current drawing would have located in ninety seconds. On a campus you service repeatedly, that is your own hours.

A call that spent forty minutes on a drawing

A municipal office building, eleven years after a mechanical renovation. The call is a leaking coil connection on a fan-coil serving a second-floor wing, reported by an occupant, dispatched through the work-order system with the room number and nothing else.

Before touching anything, the technician needs the isolation point for that branch so the coil can be worked without draining a riser that serves the whole wing. The facilities office produces a drawing set from the renovation. It shows two branch isolation valves in the second-floor mechanical closet.

Neither valve is there. What is there is a single valve, larger than either one drawn, in a different position on the wall, with a newer tag number that appears nowhere in the set.

The technician spends forty minutes on this: fifteen tracing pipe visually through an accessible ceiling, ten with the in-house maintenance lead who remembers "some work up there a few years back," and fifteen confirming by opening two ceiling tiles that the single valve does in fact isolate the branch and not more. Before closing it, he asks the facilities lead which rooms are downstream and gets the wing confirmed, because closing an isolation valve in an occupied building takes conditioning away from whoever is behind it and a valve that isolates more than you think turns a one-room repair into a building complaint. He also checks where the drain point discharges and routes it to a floor drain rather than the mechanical closet floor, because that closet has a panel at floor level and putting water somewhere it was not before is a hazard he is creating, not one the building handed him.

The repair itself takes about an hour. The locating took forty minutes on top of it.

Where the drift came from

Nobody was negligent. Reconstructing it with the maintenance lead, three separate events each moved the building away from the drawing:

Year 2, a leak repair. A section of branch pipe failed at a fitting. The repairing contractor replaced the run and consolidated two branch valves into one because the second one was redundant to how the wing was actually zoned. Sensible work. No drawing existed on site to mark, and no one asked for one afterward.

Year 6, a partition change. The building was re-partitioned for a department move. Two rooms became one, the ceiling grid was reworked, and access to what had been the second valve location was covered by new hard ceiling. The architectural drawings were updated for the partition change, because that work required a permit. The mechanical set was not, because nobody touched mechanical.

Year 9, a controls upgrade. New tag numbers were assigned campus-wide under a numbering scheme the facilities department adopted after the renovation. The valve now carries a tag from the new scheme. The drawing carries the old one. Both are correct within their own system, and there is no cross-reference between them.

Notice the shape. One change was made by a contractor with nowhere to record it, one was made in a discipline that did not think it affected another discipline, and one was an administrative renumbering that touched every asset at once. Only the third was intentional.

The drift is asymmetric, and that is why it compounds

Two properties make this worse over time rather than staying flat.

It only moves one direction. Nothing in normal operation moves a building back toward its drawings. Every undocumented change adds to the gap, and none subtract. A set that was 95% accurate at handover does not oscillate around 95%; it decays.

Later changes are harder to detect than earlier ones. The first undocumented change leaves visible evidence, such as newer pipe against older. By the third, the newer work has itself aged, and the technician has no way to tell which of the things in front of them is original. That is why the forty minutes above went to tracing rather than to a quick comparison.

The direction check that keeps this honest: at the other end of the range, a building whose record set was updated at every touch gives you locating time close to walking time, a couple of minutes, because you read the drawing and go. The forty minutes is not a fixed cost of institutional work. It is the interest on eleven years of unrecorded changes.

Three documents, commonly confused

Most shops produce the first and deliver the second. Almost no institution maintains the third well, and that is the gap. An as-built handed over at closeout is a snapshot with a date on it, accurate the day it was made and decaying from that day. A record drawing is a process, and the process usually has no owner.

The three-minute habit

The habit that changes this is not a drafting program. It is a photograph plus two written lines, taken while the change is in front of you:

1. Photograph the change with a fixed reference in frame. A door frame, a column line, a beam, a labeled panel. A photograph of a valve against blank pipe tells the next person nothing about where it is. With a tape measure to a reference, it locates the valve.
2. Write the two lines that a photo cannot carry: what was there before, and what it serves now. "Two branch valves consolidated to one; this valve isolates rooms 204 through 214" is the whole record.
3. Send it to the owner's record, not just to your own job file. This is the step shops skip and it is the one that matters, covered below.

Three minutes at the moment of the change. There is no version of reconstructing it later that costs less.

Worked example: what eleven years of drift cost, in hours

Pull the shop's own service log for that one mechanical room over the four visits it made in three years. Time spent locating isolation or tracing routing before productive work began:

• Visit 1, coil leak: 40 minutes
• Visit 2, control valve replacement: 25 minutes
• Visit 3, riser investigation after a complaint: 55 minutes
• Visit 4, filter and belt service on the adjacent air handler: 20 minutes

Locating total: 140 minutes, which is 2.33 hours. Productive wrench time across the same four visits: 9.5 hours. So locating ran at about 25% of productive time, and about 20% of the 11.83 hours of total on-site time across those visits. State whichever base you use, because the two numbers describe the same waste and a reader who mixes them will quote the wrong one.

Visit 3 is the outlier and it is worth naming as one rather than averaging it away: 55 minutes is more than double visit 4, because visit 3 was an investigation with no starting point rather than a scheduled task at a known location. Investigations are where drift costs the most, which means the shops that mostly do scheduled maintenance underestimate the problem and the shops that get called for the odd complaint feel all of it.

Now the other side. Suppose the shop had spent 3 minutes redlining at each of the four visits, 12 minutes total, and each redline had reached the owner's record. Visit 2 onward would have started from a current drawing. Conservatively, that removes most of the locating time on visits 2 through 4, roughly 100 minutes of the 140, at a cost of 12 minutes. That is not a rounding difference; it is a ratio of about eight to one in time recovered against time spent, and it is time recovered on your future visits rather than a favor to the owner. There is a third limit on that ratio and shops raise it first: on straight time and materials those 140 minutes are billed, so the redline gives away revenue rather than recovering it. The eight to one holds on fixed price, on a not-to-exceed band, and anywhere your number is compared against a competitor's. It does not hold on open time and materials, though neither does the account, once the director notices that a third of the invoice was spent looking for a valve.

Two honest limits on that figure. First, it only holds where the same shop keeps returning to the same building, which is exactly the institutional case and is not true of one-off residential work. Second, it assumes the redline actually reaches a record the next technician opens; a redline that dies in your own job folder buys you nothing on a visit run by a different crew.

Your copy does not count

The single most common failure is a shop that redlines diligently into its own job files. When the next call comes in eighteen months, the dispatcher does not know the file exists, the technician does not open it, and the drawing the facilities office hands over is still the old one.

Make the destination explicit:

• Attach the markup to the work order in the owner's system, if their work-order or CMMS platform accepts attachments. That is where the next person looks, because it is where the building's history lives, and on institutional accounts the ticket frequently outlives every other record of the visit.
• Ask who maintains the record set and send it there by name. In most facilities departments this is one person, often the same person who holds the campus standards list.
• Ask whether they want redlines at all. Some institutions have a formal process and will reject informal markups; some have nothing and will be grateful. Both answers are useful, and asking positions you as the vendor who noticed.
• Keep your own copy too, because the owner's process may not survive a staff change.

Where this reads differently

On leased space with a property manager. A property manager administers someone else's asset and may have no record set at all, only a lease file. The drift is the same but the destination for your redline is the ownership entity or its engineer rather than an in-house facilities department, and it is worth asking who that is before assuming your markup is going anywhere.

On a campus with an active capital program. Where an institution is designing a renovation, an accurate existing-condition record is worth far more than usual, because the design team is about to pay for a field survey. A shop that can hand over eleven years of documented changes to the building is handing over something the design team was going to buy.

On equipment rather than distribution. Nameplate and setpoint records drift the same way, but they are recoverable by walking to the unit. Concealed distribution is the case where the record is the only source, because the alternative is opening a ceiling.

How to verify you got this right

• Pick one building you service regularly and time the pre-work portion of your next three visits there separately from productive time. If locating is above roughly 10% of on-site hours across those visits, drift is charging you and you have the number to say so.
• Open your last five job folders and check whether any redline in them ever left your office. If none did, you have the habit and not the benefit.
• Ask your facilities contact to show you the record set for one system. What they hand over, and how quickly, tells you which of the three documents above the institution actually maintains.

References

• See related: How to Assemble a Closeout Package, How to Run a First Walkthrough on an Institutional Site, How to Close Out a Work Order in the Field
• Trade-standard practice for record drawings and field markups in construction closeout
• International Facility Management Association, general guidance on building documentation and asset records