Construction commissioning is the process that proves the building systems do what they were designed to do. It sits across the design contract, the construction contract and often a separate commissioning agreement, and the gaps between those three surface in the last eight weeks of a project, which is the worst possible moment.
Construction commissioning has four levels, and they sit apart
The word covers several distinct activities and contracts frequently fail to distinguish them.
Factory testing. Equipment tested at the manufacturer before shipping. Almost always in the equipment supply price, occasionally with owner witnessing as an extra.
Startup. Getting each piece of equipment running. Clearly the installing trade responsibility and rarely disputed.
Functional testing. Proving that systems perform their sequences, including under failure conditions. This is where the arguments are.
Integrated systems testing. Proving that multiple systems work together. Fire alarm triggering smoke control, triggering elevator recall, triggering door release, on emergency power. The most valuable test and the one most often omitted.
The first two are almost always covered. The third is partly covered and partly assumed. The fourth is frequently in nobody price at all, because it belongs to no single trade and requires all of them present at once.
The three contracts and the gaps between them
On a typical project, commissioning responsibilities are divided across three agreements that were negotiated separately.
The design contract may include development of a commissioning plan, review of submittals against design intent, and witnessing. Frequently it includes the first and not the others.
The construction contract includes startup and usually functional testing by each trade, and the general contractor coordination of the process.
A separate commissioning agent, where one exists, is engaged either by the owner or by the contractor, and that distinction matters more than anything else in this subject.
The gaps appear in four places. Who writes the test scripts. Who provides temporary power, water and instrumentation for testing. Who pays for the trades to stand present during integrated testing. And who pays to re test when a test fails.
That last one is the expensive gap, because integrated tests fail routinely on first attempt, and a project with no re test provision has each failure becoming a commercial discussion at a point when the opening date is fixed.
Why the commissioning agent reporting line decides the outcome
A commissioning agent engaged by the general contractor is reviewing the work of the party paying it.
That is not an accusation of bad faith. Competent agents work that way on many projects and produce honest reports. But the incentive on a difficult call, at a moment when the project needs to open, runs in one direction.
An agent engaged by the owner reports to the owner, and is not commercially dependent on the outcome. The cost difference is small, typically well under one percent of construction cost, and the position is substantially different.
There is a second consequence. An owner engaged agent can be brought in during design, reviewing the sequences and the testability of the systems while changes are still cheap. A contractor engaged agent usually arrives after the contract is signed and reviews what already exists.
Where the agent sits inside the construction contract, the useful mitigations are a right of direct communication with the owner, a requirement that all reports go to both parties simultaneously, and a stated scope that cannot be reduced without owner consent.
A worked example
Illustrative figures. Not taken from any client project and not a quotation.
A 96,000 square foot medical office building with a fixed opening date driven by a tenant lease.
Commissioning is carried in the construction contract as a general obligation to test and demonstrate systems. No commissioning agent is separately engaged. No integrated systems test is specified.
Individual systems test satisfactorily. Six weeks before opening, the first attempt at an integrated life safety test fails: smoke control does not achieve the required pressure differentials with the doors in their tested configuration, and the sequence between the fire alarm and the elevator recall does not perform as designed.
Resolution takes eleven weeks. Duct modifications $180,000, additional controls work and reprogramming $210,000, trades standing time through five re tests $140,000, and consultant engagement $60,000. Plus an opening five weeks late with $90,000 of associated cost.
Total $680,000.
A specified integrated test with an agreed script, run twelve weeks earlier as a first attempt rather than as an acceptance test, would have found the same two failures with time to correct them inside the ordinary sequence.
What to specify before signature
Five items, each short, and together they cover most of what goes wrong.
The commissioning scope by level, saying explicitly which of the four levels is included and for which systems.
Who writes the test scripts and when they are issued. Scripts written by the party being tested are a weaker instrument than scripts written by the design team or the agent.
Trade attendance at integrated testing, named as an obligation in each trade scope rather than assumed. This is the item that most often becomes a change order.
Re testing, stating how many attempts are included and who pays beyond that. Two included attempts is a reasonable position and it changes the preparation.
A commissioning schedule with its own dates, integrated into the construction schedule rather than compressed into the final month. Where testing sits on the critical path, it belongs in the analysis described in whether the completion date is real.
The connection to substantial completion
Commissioning is also the practical test of whether a building is genuinely complete, which is why it belongs in the completion definition rather than alongside it.
A building where every system is installed and no system has been proven is physically finished and operationally unknown. An owner accepting it has accepted the risk of every system that does not work.
Requiring commissioning reports as a condition of substantial completion changes that, for the reasons set out in substantial completion and the punch list.
The objection raised is that this delays completion and therefore costs the owner money in extended financing. Sometimes true. What it costs more reliably is the alternative: a year of chasing trades who have been paid and have demobilized, to fix systems that were never demonstrated.
The second connection is to the warranty. Warranty periods start at substantial completion, and a system that is not commissioned until three months later has consumed a quarter of its warranty before anybody has used it.
There is a third connection, to the operator. On a hotel, a hospital or any building with a facilities team, the people who will run the systems should be present during testing rather than trained afterwards from a manual. That attendance is a contract requirement or it does not happen.
Owners who arrange it find that the first year of operation is materially easier, for a cost of nothing beyond a sentence in the scope.
The last point concerns the reports themselves. A commissioning program produces a substantial record: test scripts, results, deficiencies, retests and a final report. That record is the evidence base for every warranty claim in the following two years, and it should be delivered as a condition of final payment rather than requested afterwards.
Owners who take a building without it are relying on the memory of people who have moved to another project.
What we do
We read the commissioning scope across all three contracts to find what is in none of them, specify the integrated test and the re test position, and set the commissioning dates against the construction schedule rather than at the end of it. Where the agent sits with the contractor, we set the reporting requirements that make the reports useful. The work is part of the constructability and interface review. With three weeks or more before signature, the full pre-GMP review reads the price, the schedule, the interfaces and the change exposure together.
Questions people ask
Is a commissioning agent worth the cost?
On any building with integrated life safety, complex controls or a fixed opening date, almost always. The fee is typically a fraction of a percent of construction cost and the exposure it addresses is measured in weeks of delay. On a simple warehouse the calculation is different and often goes the other way.
When should commissioning planning start?
During design, because that is when sequences can still be made testable and when access for testing can be designed in. A plan written after the contract is signed inherits whatever was drawn, including systems that cannot be tested without modifications nobody priced.
What if the contractor says commissioning is included?
Ask which of the four levels, for which systems, with how many re tests, and who attends. Included is not a scope. The answers to those four questions either describe a real program or reveal that startup has been priced and functional testing has been assumed.
This is general information about construction contracts and is not legal advice.