Every bookkeeper hits this moment. You're drowning in payment posting, spending hours matching remittance PDFs to QBO invoices, and somebody says: "Why don't you just hire a VA?" It sounds reasonable. Labor costs consume 62% of total AP processing costs (APQC via ResolvePay, 2025). If labor is the biggest expense, throw more labor at it. Problem solved.
Except it isn't. I wrote about the payment matching problem a few weeks ago. This is the other side of that story - why I chose to build instead of hire.
TL;DR: Hiring a VA for payment posting sounds logical, but it doesn't fix the workflow. Manual invoice processing costs $12-$16 per invoice (IOFM via ResolvePay, 2025), and that cost follows you regardless of who's clicking the buttons. Building automation solved the problem once for every client, instead of retraining for each one.
How did payment posting become my biggest time sink?
AP professionals spend more than 10 hours per week processing invoices, according to DocuClipper (2025, citing Medius). I wasn't tracking my hours that precisely, but when I finally did, the number was worse than I expected. Payment posting was eating my weeks alive.
It crept up on me. With three clients, posting payments was just part of the routine. Open the remittance PDF, open QBO, match invoice numbers, enter amounts, save. Fifteen minutes here, twenty minutes there. Annoying but manageable.
Then I hit eight clients. Then twelve. Each one with their own remittance format. Each one with their own quirks. One sends clean PDFs. Another sends screenshots from their bank app. A third types invoice numbers into the body of an email with no formatting. You know these people. You have them too.
From my practice: The week I realized I'd spent six hours on payment posting alone, across four clients, was the week I started questioning everything. Six hours of copying numbers from documents into QBO. Not analyzing anything. Not advising anyone. Just being a human data-entry bridge between two systems that should be talking to each other.
The volume kept climbing, but the workflow never got easier. Every new client meant another set of remittance formats to learn, another batch of invoices to match manually. I was scaling my time linearly with my client count. That's a ceiling, not a business.
Why does "just hire a VA" sound right but fall apart?
The virtual assistant market is projected to exceed $19 billion by 2025 (Grand View Research, 2023). Clearly, outsourcing labor is a massive industry. And for plenty of tasks, VAs are the right call. But for payment posting specifically, the logic breaks down fast.
Training never ends
Here's the first thing nobody warns you about. Payment posting isn't one skill. It's a different skill for every client. Each client sends remittances in a different format. Each QBO file has different invoice numbering. Each industry has quirks - partial payments, short pays, credits applied.
Training a VA on Client A doesn't help with Client B. You're essentially writing a new SOP for every file. And when the client changes their format (they will), you're retraining again. I've watched bookkeepers spend more time managing a VA's payment work than it would take to just do it themselves.
Errors compound quietly
Manual invoice processing has an error rate of approximately 1.6% per invoice (ResolvePay, 2025, citing Turing IT Labs). That sounds small. But a VA who doesn't know the client's history will hit edge cases you'd catch instantly. A $200 payment applied to the wrong invoice doesn't show up until reconciliation. By then, you're spending 30 minutes reconstructing what happened.
The real issue: A VA adds labor. But the problem isn't a labor shortage. The problem is a broken workflow. Adding hands to a broken process just means more hands touching a broken process. You need fewer touches, not more.
It doesn't scale the way you think
Say you hire a VA for 10 hours a week at $15/hour. That's $600/month. You now need to manage that person, review their work, answer their questions, and fix their mistakes. And when you add your next three clients? You need more VA hours, more SOPs, more review time.
Compare that to a tool. A tool learns the extraction logic once. It doesn't forget. It doesn't need a Slack channel. It doesn't call in sick. And adding a new client means connecting one more QBO file, not writing another training document.
What's actually broken about the payment posting workflow?
Only 43% of accountants have automated data entry, despite 95% adopting some form of automation (Intuit QuickBooks, 2025). That gap tells you everything. We've automated the easy parts. The hard part - the messy middle of parsing remittances and matching them to invoices - is still entirely manual.
QBO's Receive Payment screen is built for one-at-a-time
Open QBO. Search for the customer. Click Receive Payment. Scroll through the list of open invoices. Cross-reference your remittance. Enter the amount. Save. Repeat.
That workflow was designed for a business owner recording a single payment after a customer pays. It was never designed for a bookkeeper batch-processing 40 payments from a single remittance PDF. There's no upload function. No batch mode. No way to say "here's a document with 15 payments, match them all."
If you haven't read it yet, I broke down exactly why the QBO Receive Payment workflow is still broken.
The remittance format problem
Your clients don't send you standardized data. They send PDFs, CSVs, Excel files, screenshots, photos of check stubs, and plain-text emails with invoice numbers scattered across paragraphs. Each format requires you to be a different kind of parser. And this is the part that makes VAs especially fragile - every new format is a new training session.
The real bottleneck is the translation layer
Between "client sends remittance" and "payment posts in QBO," there's an invisible layer of work. Reading the document. Identifying the relevant numbers. Matching those numbers to open invoices. Handling discrepancies. That translation layer is 90% of the effort. The actual button-clicking in QBO is trivial.
A VA handles the button-clicking. Automation handles the translation layer. That's the difference.
What happens when you solve it with software instead?
Automated AP departments process invoices at roughly 6x the throughput of manual operations, reducing per-invoice costs by 70-80% (IOFM via ResolvePay, 2025). The numbers are dramatic, but they made intuitive sense to me before I ever saw the research. I was living it.
From my practice: The first time I ran a remittance PDF through the tool I was building and watched it correctly extract 12 invoice numbers, match them to open QBO invoices, and present the whole batch for one-click approval, I felt something shift. Not excitement about the tech. Relief. This problem was actually fixable.
Solving once versus solving forever
When I built the extraction logic, I solved the PDF parsing problem for every client who sends PDFs. When I built the CSV handler, I solved it for every client who sends spreadsheets. Each capability I added worked across my entire client base immediately. No retraining. No new SOPs.
That's the fundamental difference. A VA solves your problem today, for this client, with this format. A tool solves the category of problem. Permanently.
The bookkeeper stays in the loop
I want to be clear about something. I didn't build a system that posts payments blindly. Every match goes through an approval step. You see the extracted data, the proposed QBO invoice match, and the amounts. You approve or correct. Then it posts.
The judgment is still yours. The data entry isn't. That's the line I wanted to draw.
Multi-tenant from the start
A VA working across your clients needs access to each QBO file, context for each client's habits, and separate mental models for each engagement. I built Ground Control Payments to be multi-tenant from day one. Connect your QBO files, upload remittances for any client, and the system handles the routing. Because you don't have one client. You have twenty.
How did this become Ground Control Payments?
The accounting software market is expected to reach $28.1 billion by 2027 (Fortune Business Insights, 2023). Most of that investment targets business owners or enterprise finance teams. Very little targets the independent bookkeeper. That gap is exactly where I found myself building.
I didn't set out to build a product. I set out to fix my own workflow. The tool started as a script that could read a PDF and pull out invoice numbers. Then I added QBO matching. Then approval emails. Then multi-client support. At some point, I looked at what I'd built and realized other bookkeepers were dealing with the exact same grind.
Ground Control Payments works like this: upload a PDF, CSV, or image. Forward an email. The system uses AI to extract payment details - invoice numbers, amounts, payer information. It matches those against open invoices in your connected QBO file. You get an approval screen showing every match. Click approve, and the payments post to Undeposited Funds.
It's one tool that does one job. No platform. No suite. Just the part of your workflow that shouldn't be manual anymore.
Why not just use an enterprise AR tool?
Believe me, I looked. Tools like Tesorio, HighRadius, and Billtrust are serious products. But they're priced for mid-market companies with dedicated AP departments. They assume ERP-level infrastructure. A solo bookkeeper managing fifteen QBO files isn't their customer. I needed something built for how we actually work - messy inputs, multiple clients, QBO as the system of record.
Should you build or hire? A framework that actually helps
Not every bookkeeper should build software. Obviously. But the question of "build versus hire" applies more broadly than you might think. Here's how I'd frame the decision for any workflow bottleneck.
Hire a VA when the problem is volume on a stable process
If you have a clean, consistent workflow and you just need more hands to do it, a VA is the right call. Transaction categorization with clear rules. Bank reconciliation on a well-maintained file. Follow-ups with clients who owe documents. These are defined processes where adding a competent person genuinely helps.
Automate when the problem is the process itself
If the bottleneck involves translating between formats, matching across systems, or handling inconsistent inputs, automation wins. The messier and more variable the input, the worse a VA will perform relative to a well-built tool. Payment posting sits firmly in this category.
The hybrid model
Here's what I actually do now. Ground Control Payments handles the extraction, matching, and posting. I review the approvals. For everything that doesn't fit neatly into automation - client communications, judgment calls on weird transactions, advisory work - that's where human time belongs. Yours or a VA's.
The goal isn't to eliminate people. It's to stop using people as data-entry bridges.
Frequently asked questions
How much does a VA for bookkeeping typically cost?
Bookkeeping VAs typically run $12-$25/hour depending on experience and location, with many firms spending $500-$1,500/month for part-time support. But the true cost includes your management time, error correction, and retraining when processes change. When 62% of AP costs are labor (APQC, 2025), adding more labor at a lower rate doesn't change the cost structure as much as you'd expect.
Can't a well-trained VA be just as accurate as software?
A skilled VA can absolutely be accurate. But manual processing carries a baseline error rate of about 1.6% per invoice (ResolvePay, 2025). That rate holds regardless of who's doing the manual work. The issue isn't skill - it's the nature of manual data entry across inconsistent formats. Automation removes the data-entry step entirely, which eliminates that category of error.
Is Ground Control Payments only for payment posting?
Yes. It does one thing: get payments out of remittance documents and into QBO. Upload a PDF, CSV, or image. The AI extracts invoice numbers and amounts, matches them to open QBO invoices, and you approve with one click. It's built by a bookkeeper for bookkeepers, and it's intentionally narrow. One tool, one job, done well.
You can see how it works and sign up here.
What if I already have a VA and want to keep them?
Keep them. Automation and a VA aren't mutually exclusive. Let the tool handle extraction and matching - the repetitive, error-prone part. Let your VA handle the work that actually requires a human: client follow-ups, document chasing, categorization, and review. You'll get better output from both.
The bottom line
I didn't build Ground Control Payments because I'm a software developer who thought bookkeeping looked like a good market. I built it because I'm a bookkeeper who got tired of the same manual grind every week. The "hire a VA" advice kept coming, and it kept not solving the actual problem.
The problem was never that I needed more hands. It was that the workflow itself was broken. QBO doesn't handle batch payment posting. Remittances come in every format imaginable. And the translation layer between document and ledger was eating my practice alive.
If that sounds familiar, take a look at Ground Control Payments. It's live, and bookkeepers are using it. One said: "This is amazing... posted one payment against all of these invoices with the click of a button." It might save you from having the same "should I hire a VA?" conversation for the fifth time.
Or just keep copying numbers from PDFs. I won't judge. I did it for years.