10 Hours Back Per Batch
10 Hours Back Per Batch
The real cost of processing invoices by hand, and what happens when you stop
Nobody gets excited about accounts payable. It’s the back office of the back office. Invoices come in, somebody enters them, somebody else approves them, somebody prints a check, stamps it, seals it, mails it. It’s boring, it works, and that’s exactly why it’s the last thing anyone bothers to fix.
Which is how you end up spending 10 hours on a single AP batch and not even realizing that’s abnormal.
Hospital Association of Oregon: 3.5 years in
The Hospital Association of Oregon — a mid-size nonprofit running advocacy and services for the state’s hospitals — was doing AP the way most companies still do it. Invoices arrived as email attachments. Approvals happened over email. Checks got printed, physically stamped, sealed in envelopes, and mailed. Payment vouchers had to be hand-keyed into their accounting system twice because nothing talked to anything else.
Jason Hershey, their VP of Finance and Accounting, put the number on it: “Our previous bill pay process probably took a good 10 hours per AP batch. Now it just takes a couple of minutes between getting an invoice entered, approved, and processed.”
Ten hours to a couple minutes. They implemented Ramp Bill Pay, connected it to their NetSuite accounting system, and the results over 3.5 years tell the whole story: 700+ hours reclaimed by the finance team, 17 hours per month freed up, month-end close shortened by 5 days, and 92% of their AP process now runs automated. Less than 10% of invoices get touched by a human.
That’s a real company with a real quote from a real person on the record. Not a vendor projection.
Now run the math on your shop
The benchmarks on manual invoice processing are well-documented at this point. Quadient’s 2025 AP statistics put the average cost at $12.88 per invoice, with less efficient shops running up to $19.83. IOFM — the Institute of Finance and Management, the people who literally benchmark this stuff for a living — puts the range at $10-$23 per invoice depending on complexity. Best-in-class automated teams? $2.78 per invoice.
Let’s take a plumbing company with 40 employees. They’re processing maybe 400 invoices a month — vendor supplies, subcontractors, equipment leases, fuel cards, insurance. Nothing exotic. Just the normal grind of running a trades business.
At $12.88 per invoice, that’s $5,152 a month in processing costs. Call it $61,800 a year. Most of that is labor: someone’s time spent on data entry, chasing approvals, cutting checks, filing paperwork. It’s not a line item anyone tracks because it’s spread across two or three people who also do other things. But it’s real money leaving the building every month.
Drop that to $2.78 per invoice with automation and you’re at $1,112 a month. $13,344 a year. The delta is $48,456 annually. That’s a full-time employee’s salary being absorbed by a process that could run itself.
And the time numbers are just as ugly. Quadient’s data shows average AP departments take 9.2 days to process a single invoice. Best-in-class automated teams do it in 3.1 days. Fully automated workflows process 30 invoices per hour versus 5 manually — a 6x throughput improvement. For a company processing 400 invoices a month, that’s the difference between someone spending 80 hours on invoice processing or 13.
Why nobody fixes this
AP processing is invisible overhead. It doesn’t show up as a single budget line — it’s distributed across payroll, office supplies, and “admin time.” Nobody’s tracking it because nobody’s measuring it.
It also works. Barely, inefficiently, with occasional duplicate payments and late fees — but it works. And “it works” is the most powerful force in business inertia.
The bookkeeper and the office manager have been absorbing this for years. They’ve built workarounds. Nobody with budget authority is watching.
The Hospital Association of Oregon had all three of these problems. Email-based approvals, paper vouchers, manual check runs, double data entry. Every one of those is a process that someone had been doing long enough to stop questioning it.
What the numbers actually mean
AP automation is one of the most boring, most proven, least risky places to start with operational technology improvements. There’s no AI hallucination risk. There’s no “will the model be accurate” question. It’s OCR reading invoice numbers, matching them to POs, routing approvals, and scheduling payments. The failure modes are well-understood and the benchmarks are a decade deep.
Sixty-two percent of organizations adopting AP automation say they did it to reduce manual errors and improve speed. Seventy-four percent of AP departments plan to use AI in their processes.
None of that matters to the plumbing company owner. What matters is this: there’s almost certainly a process inside your operation that’s eating 10 hours nobody’s counting, costing $4-5K a month nobody’s tracking, and producing errors nobody’s measuring. AP happens to be the one with the best benchmarks. But the pattern repeats everywhere — in scheduling, in estimating, in customer intake, in inventory reconciliation.
The Hospital Association of Oregon connected software to software and stopped printing checks. The 700 hours they got back were just hours that never should have been lost in the first place.