Architecture job postings are oddly revealing documents. Read enough of them and you start to see the operating model peeking through.
A project coordinator listing will often describe the role as the engine that keeps projects moving. Then the responsibilities start stacking up: submittal tracking, RFI routing, pay application review support, field report distribution, current-set updates, meeting notes, follow-up. The posting sounds like headcount planning. Most of the time it is really process leakage, written out as a job description.
And that raises the uncomfortable question: how much of this role is judgment work, and how much of it is repeatable coordination that the firm has learned to handle with people?
The Hidden Math Of Construction Administration
Most construction administration roles contain both kinds of work. That part is normal. What gets expensive is pretending they are all the same kind of work.
A minority of the role requires professional judgment. Reviewing pay applications for compliance. Interpreting RFIs. Catching errors in contractor submissions. You want an architect or experienced reviewer making those calls.
Much more of the role is movement. Routing documents. Chasing status. Updating logs. Sending the approved thing to the next person. Maintaining a current set. This work is real. It matters. Projects get messy fast when it is neglected. But it is not the same as judgment.
Then there is the overhead layer: training, backup coverage, context loss when someone leaves, and the slow rebuilding of firm-specific knowledge when the next person inherits the seat.
That is the hidden math. Every additional coordinator can absorb work volume, but each one also increases the amount of coordination your coordination system requires. Someone has to train them. Someone has to cover when they are out. Someone has to reconstruct the local logic of how the firm actually moves information through projects.
What Infrastructure Changes
I am not arguing that architecture firms should stop hiring project coordinators. I am arguing that firms should get much more honest about which parts of the role are really operating-system work.
Take a submittal. In a lot of firms it lands in a shared inbox, someone triages it manually, forwards it to the reviewer, follows up for status, updates a spreadsheet, and then notifies the contractor once the item is approved.
None of that is trivial. But most of it is also not design judgment.
A workflow layer can hold that movement. The system can identify the responsible reviewer, open the tracking item, keep the status current, and route the output where it needs to go. The architect still decides whether the submission is acceptable. The review stays human. The movement around the review stops depending on memory and discipline alone.
That same logic often applies to RFIs, field report distribution, current-set updates, meeting note circulation, and other construction administration chores that are painfully important and painfully repetitive.
The Three-Workflow Test
Before you post the next coordinator role, run a simple test.
List the repeating tasks the person would handle every week. Then mark the ones that follow a predictable pattern. Then ask which of those patterns could be enforced by workflow instead of social pressure.
If three or more of the recurring tasks fall into that bucket, you may not be looking at a staffing problem yet. You may be looking at an operations problem you have been solving with headcount.
Where Hiring Still Makes Sense
There are absolutely cases where adding a person is the right move.
If the coordination load really is full of project-specific exceptions, you need someone with judgment, not infrastructure.
If the firm is already standardizing the repetitive work and now needs an owner for the workflows, that is also a real role.
If growth is happening fast enough that you need both better infrastructure and a person responsible for improving it, fine. That is a different conversation from “we are buried in repetitive coordination and need another set of hands.”
The distinction matters because the long-term costs are different. Headcount is recurring. Process inconsistency is recurring. Knowledge loss is recurring. A workflow layer is front-loaded, but it tends to get cheaper per project as volume rises and the team stops reinventing the same movement every week.
The Better Question
If I were looking at a draft project coordinator posting, I would ask one question before anything else: of the responsibilities listed here, which ones require professional judgment, and which ones require follow-through?
That split will usually tell you more than the job title does.
If most of the role is follow-through, I would pause before posting it. You may still decide to hire. But you should know what you are buying. A person can absolutely carry an underbuilt operating system for a while. Firms do it every day. It is just an expensive way to keep the work moving.
Your most valuable people should spend their time on judgment. The rest of the movement deserves better infrastructure than crossed fingers and a shared inbox.
If you’re navigating this decision, I offer a First AI Workflow Audit that identifies the highest-value coordination workflows in your construction administration process. I’ll map what requires human judgment and what could be infrastructured. Takes 90 minutes. Get in touch if you want to walk through it.