A rooftop unit fails on a ninety-degree afternoon. The store manager texts the regional facilities lead. The lead calls the contractor. A technician quotes a compressor replacement over the phone, and the lead approves it verbally because the site is losing business by the hour.
Three weeks later, an invoice lands in accounts payable that nobody can verify.
Shawn Grant, founder and CEO of ClearFM, has watched this sequence play out across multi-site portfolios for years, from both sides. He ran the service work getting phoned in and handled fast. He also sat with the invoice weeks later, reconstructing what was agreed from memory and scattered text threads. His conclusion is structural.
“Every person in that chain did their job. The operating record fell apart at each handoff.”
Urgency Exposes the Design
The easy read on a missing record is that someone got sloppy under pressure. Grant doesn’t buy that, and the evidence backs him up.
Nobody’s thinking about documentation when a unit is down and tenants are calling. That’s a reasonable human response to a real emergency. The coordinator who approves a repair verbally is protecting the business in the only way available at that moment.
The problem is in the operating model. Facility workflows are designed for the calm path, where a request gets entered, routed, approved, and closed in sequence. Emergencies get treated as exceptions, even though any multi-site portfolio produces them constantly.
A model that only holds together when nobody’s rushing is going to have blind spots. The rooftop unit scenario repeats because the conditions that produce it repeat.
Why the Gap Surfaces in Accounts Payable
Grant traces this in his executive white paper on facility operations. The request lives in a text thread. The approval lives in a phone call. The governing contract sits in a folder nobody opened. The completion photos sit on a technician’s phone. It’s scattered across four places before anyone’s asked to write a check.
None of this looks broken while the work is happening. The unit gets fixed. Everyone moves on.
That becomes visible weeks later, when the work turns into money. An AP reviewer gets a PDF invoice that references none of the context, and the amount runs roughly double what the regional lead remembers. The reviewer now has two bad options: approve on trust, or slow the payment down with email interrogations of facilities.
Accounts payable is where the missing record gets discovered because payment is the first moment someone has to prove what was agreed. The failure happened weeks earlier, in motion, at the handoffs. The financial system just reports it last.
Design for the Moment It Breaks
Grant looks at operating models by how they behave under stress. His test is simple: when everyone’s moving fast, does a path exist to capture what mattered, and does that path cost less effort than the workaround? It’s not a sophisticated test. Most organizations fail it.
A few things follow from that.
Give emergencies an expedited lane, with a record attached. If the standard process can’t handle urgent work, people route around it. The workaround becomes the actual operating model. An expedited path with a named approver and a captured decision keeps speed and evidence together.
Make evidence travel with the work. The approval should reference the contract rate. The invoice should reference the approved scope and the completion evidence. When context moves with the job, the AP reviewer compares instead of reconstructing. That’s the part most systems skip.
Pay attention to the workarounds. Coordinators and site staff can name every shortcut in the current process. Those shortcuts mark exactly where the official path breaks under pressure. Honest diagnostic, if anyone asks them.
What a Healthy Portfolio Looks Like at Invoice Time
Grant looks for a specific signal when he’s assessing a portfolio: no scrambling at invoice time. When AP can trace a payment to an approved scope in a few minutes, the upstream handoffs held. When every emergency produces a dispute, the model’s running on memory and everyone’s acting like that’s fine.
For anyone accountable for facility spend across multiple sites: the next rooftop unit will fail on a hot afternoon, and your teams will move fast, because that’s their job.
Whether the record survives that afternoon depends on decisions made now, in calm conditions, about how handoffs have to work when nobody has time to think about them.
The emergency is the recurring stress test. Build for it, and the invoice that arrives three weeks later is either a lookup or an investigation. That part’s already decided by decisions made long before the unit fails.