The In-House Maintenance Staff as a Peer, Not an Obstacle
Why this matters
The in-house crew is the only party on an institutional site who can end your relationship without anyone above them noticing. They do not sign your purchase order and they do not score your bid, but they decide whether your technician waits twenty minutes for a key or two hours, whether the undocumented valve change from six years ago gets mentioned before you chase it, and what the director hears when your name comes up. Shops treat them as gatekeepers to be worked around. That is both wrong and slow. The more useful frame is that you and they are running one maintenance program with a boundary in it, and the boundary needs to be written down by work type rather than negotiated per incident.
Who they actually are
An institutional maintenance crew is usually multi-trade by necessity. A stationary engineer or lead mechanic, one or two general maintenance technicians, sometimes a licensed electrician or a licensed plumber depending on the state and the institution's size, plus custodial or grounds staff who hold the keys to more of the building than anyone realizes.
They are almost always more capable than an outside shop assumes and almost always narrower in permitted scope than their capability suggests. The lead mechanic who has run a building's boiler plant for eighteen years knows things about that plant you will not learn in three years of service calls. That same mechanic may be prohibited from touching the burner's gas train, not because he cannot, but because of what governs the boundary.
What actually sets the boundary, and what does not
Capability does not set it. Four things do, and they operate independently, so a task can be off-limits to the in-house crew for one reason and on-limits for the other three.
License scope. State licensing determines who may perform certain electrical, plumbing, gas, refrigerant and boiler work, and an institution's own maintenance staff are not automatically exempt. Some states carve out maintenance-of-owned-property exemptions, some do not, and the carve-out often stops at a defined scope such as like-for-like replacement. Ask which exemption the institution believes it operates under, because it drives which tickets come to you.
Insurance and liability. An institution's risk manager may bar in-house staff from work classes regardless of licensing, typically anything energized, anything at height beyond a defined level, anything involving hot work, and anything on a system tied to life safety.
The collective bargaining agreement. Where the in-house crew is represented, the agreement often defines which work is theirs and which may be contracted out. This is a real constraint and it is not personal. A steward raising a jurisdiction question about work you were handed is doing their job, and the answer comes from the institution's labor relations function, never from you.
Manufacturer or warranty requirement. Equipment still inside a manufacturer's warranty period frequently requires a factory-authorized servicer, which pushes work outside regardless of who is competent to do it.
Say which of the four applies when you write a scope. "In-house crew handles filter changes" is an assumption. "In-house crew handles filter changes and belt replacement; outside shop handles all work on the gas train per the institution's hot work and licensing policy" is a boundary.
The artifact: a work-type split you can hand to both crews
This is the thing worth building on your second visit. One page, per system, by work type. Filled in below for a university central plant and its air handling equipment, with a crew of six covering eleven buildings.
| Work type | In-house | Outside shop | Why the boundary sits here |
|---|---|---|---|
| Filter changes, belt tension, lubrication | Yes | No | Routine, no license gate, in-house has the access |
| Coil cleaning, drain pan clearing | Yes | Overflow only | Capacity, not capability; outside covers the summer backlog |
| VFD parameter changes | No | Yes | Risk manager bars energized panel work by staff |
| Motor replacement, disconnect open and locked | Shared | Shared | Either crew, whoever holds the lock, never both without coordination |
| Burner and gas train service | No | Yes | Licensing plus the institution's hot work policy |
| Refrigerant recovery and charging | No | Yes | Section 608 certification held only by the outside shop |
| Boiler annual certification support | Yes | Yes | In-house preps and isolates, outside performs, both present |
| Controls programming above the graphic layer | No | Yes | Vendor-held source and licensing |
| After-hours no-heat first response | Yes | Escalation only | In-house triages, calls out only past a defined condition |
Two parts of the sheet do the real work, and neither is the yes or no. The "why" column is what makes it survive a personnel change, because a boundary with no stated reason gets relitigated by every new hire on both sides. The shared rows are where every coordination failure actually happens, which is the next section.
The shared rows are a lockout boundary, not just a scheduling one
When two crews can both work the same equipment, competence on both sides is not sufficient. Each crew's protection has to be independent of the other crew's good intentions.
Where an outside servicing employer and the institution's own staff are both involved with the same machine, 29 CFR 1910.147(f)(2) requires each employer to inform the other of their respective energy control procedures, and each authorized person applies their own lock, which is what 1910.147(f)(3) group lockout is built for. Practically: your technician does not work under the in-house mechanic's lock, and the in-house mechanic does not work under yours. Every person exposed has a device on the isolating means, and the last device off is the last person out.
For the electrical hazard specifically, that is 29 CFR 1910.333(b)(2) rather than 1910.147, which excludes exposure to electrical hazards from work on conductors and equipment in electric utilization installations at (a)(1)(ii)(C), and the conductors are proved dead with an instrument verified on a known source before and after the test, per NFPA 70E-2021, 120.5. Do not recite the regulation at them. Walk to the isolating means with your own lock in hand, hang it beside theirs, and say that yours goes on and theirs stays until you are clear. That is the same statement and it does not sound like a class.
Where the shared work involves a pit, a vault, a tank or a tunnel, the coordination duty is statutory the other direction as well: under 29 CFR 1910.146, at paragraphs (c)(8) and (c)(9), the host employer must inform the outside contractor of permit-required confined spaces and their hazards, and both parties coordinate entry operations. If nobody has told you which spaces on the site are permit-required, ask before your first entry rather than after.
What they control that you cannot buy
Access. Keys, escort availability, roof hatch codes, elevator service mode, the one door that is chained from the inside.
Undocumented history. Every institutional building carries a layer of modifications made by people who left. The isolation valve labeled for one wing that actually feeds two. The circuit that was re-fed during a renovation and never relabeled. A pump curve that has not matched the impeller since a rebuild. None of this is in the drawings. All of it is in the lead mechanic's head, and it is offered to peers and withheld from people who talk past them.
Credibility on escalation. When you tell the director a system needs replacement, the director asks their own staff whether that is true. Your recommendation is evaluated by the person you either included or bypassed.
Two ways shops burn this, both avoidable
Diagnosing over their head. A technician finds a fault the in-house crew missed, and the finding goes straight up to the director in writing. The finding may be correct. The delivery makes the in-house crew look incompetent in front of the person who scores them, and it costs you the history channel permanently. Tell the mechanic first, in person, then write it up with their observation included.
Quietly absorbing their work. A technician does a filter change because they were standing there. It reads as helpful once and as scope creep by the fourth time, and where a collective bargaining agreement is in play it can generate a grievance against the institution that your customer then has to answer. Do the work if asked, note it on the ticket, and raise the pattern with the director rather than letting it accumulate.
Worked example: a boiler certification at an eleven-building campus
The crew of six handles roughly 1,400 tickets a year. They close about 1,190 of them in house, which is 85 percent, and roughly 210, or 15 percent, go outside. Your shop is one of the vendors on the 15 percent.
The annual boiler certification comes up. Under the split sheet: in-house preps and isolates, outside performs, both present. Here is how the hours actually land, and the point of counting them is that the sheet changed who spends them.
In-house: 3.0 hours over two days ahead of the visit, draining and isolating, verifying the gas cock position, hanging their own lock on the isolating means, and pulling the prior year's certificate and the last two inspection reports out of the work-order system. Outside shop: 4.5 hours on the day, controls and safety device testing, low-water cutoff verification, plus 1.0 hour writing up the report against the institution's ticket number.
Total labor across both crews: 8.5 hours, of which your shop billed 5.5, or roughly two thirds. Without the sheet, the previous vendor billed 8.0 hours for the same certification because their technician did the isolation and the record-pulling himself, on the clock, on the second attempt after arriving to find the boiler still hot.
Read that correctly, though, and do not read it as a saving of 2.5 hours to the institution. The 3.0 in-house hours are real hours from a crew of six that is already carrying 1,400 tickets, so the institution moved cost rather than eliminating it. What it actually bought was a single-visit completion instead of a return trip, and a prep step performed by the people who know which isolating valve is mislabeled. Sell it as that, not as a saving, because the director can check.
The failure mode: the sheet says shared, both crews assume the other isolated, and someone opens a drain on a boiler at operating pressure. That is why the shared row is a lockout row and not just a scheduling row, and why the person exposed hangs their own device rather than trusting a verbal confirmation across two employers.
How to verify the boundary is real and not aspirational
Pick three tickets from the last quarter that touched the shared rows. For each, name who held the lock, who wrote the close-out, and who the director would have called if it had gone wrong. If any of the three has a different answer than the sheet says, the sheet is a document rather than a practice, and the next incident will land in the gap.
References
- 29 CFR 1910.147(f)(2) and (f)(3), multi-employer energy control coordination and group lockout; the electrical exclusion sits at (a)(1)(ii)(C)
- 29 CFR 1910.333(b)(2), electrical safe work practices for de-energized work; NFPA 70E-2021, 120.5 for the verification sequence
- 29 CFR 1910.146(c)(8) and (c)(9), host employer and contractor duties for permit-required confined spaces
- See related: What Changes When the Customer Has a Facilities Department; The Facilities Director and What They Are Actually Judged On