Is Accounting Dying? The Answer Splits by Task
Community threads keep asking whether accounting is dying as QuickBooks automates categorization, and the accepted answer splits the question by task.
Community threads around QuickBooks automation keep circling one question: is accounting dying? The answer most threads settle on is a 2018 essay by practice consultant Darren Root, published on Intuit’s blog for accountants. Its verdict was no, with a catch. Half the essay reads like prophecy. The other half now reads like history.
The fear behind the question
Users reach the question the same way. They watch bank feeds sort transactions unaided, receipts parse themselves, and invoices match payments without a click. Then they ask whether their role survives the software. The accepted answer’s first move is to split the question. It bundles three separate fears: that firms become obsolete, that the discipline stops mattering, and that humans stop typing. Only the last one deserves a yes.
No product version drives the worry, and none needs naming. The automation the essay described now runs across QuickBooks Online and updates quietly, release by release. The fear renews itself with each update, which is why the question never ages.
The accepted answer, in short
The essay sorts the three fears by outcome. If accounting means humans keying transactions into systems, that work was already in retreat when the essay appeared. Machine learning was categorizing routine activity with little human help, and companion tools for payables, expenses, and time tracking were linking in around it. If accounting means keeping the score, nothing is dying at all.
Assets, liabilities, revenue, and expenses stay necessary as long as anyone needs the score kept. Human oversight survives, the essay concedes, but far smaller. The growth, it concludes, sits in judgment and advice.
The years since have tested both halves
We weighed the essay against the software it anticipated, and it aged unevenly. The data-entry prediction aged well. Bank feeds suggest categories and refine those suggestions from your corrections. Rules route recurring items without a click. Receipt capture reads an image and fills in the fields. Payments match invoices on their own. What the essay forecast as a future is now the default behavior of the product.
The other prediction aged well in a direction the essay never spelled out. The review layer grew. Exceptions, duplicates, misapplied rules, and misread receipts all demand eyes now, and the software tends to flag them late. The bottleneck moved from keystrokes to judgment.
Errors also changed species. In the typing era, a mistake was a wrong number entered once. In the automation era, a mistake is a plausible category applied silently, again and again. Correcting a rule fixes the future and leaves the past untouched. That shift, more than any feature list, explains why firms lean harder on people who can audit the machine.
Habits that settle the question in practice
The practical half of the answer was thin in 2018, but the software has since filled it in. Treat automation as a draft, never as finished work. These habits cover most of what the draft gets wrong:
- Correct a wrong suggested category instead of retyping the transaction. The correction teaches the suggestion engine, so the fix compounds.
- Keep bank rules few, and name them clearly enough to audit later. A rule nobody remembers writing is a quiet error factory.
- Clear the uncategorized queue weekly. Whatever sits there is data entry the machine declined to guess at.
- Reconcile monthly, and sample the auto-matched lines, not only the unmatched ones. Silence is not the same as accuracy.
- Watch payee merges and duplicate downloads. Both are artifacts of automation, and both age badly when ignored.
The profession now sits on the review side
The essay’s split holds. Roles built around typing have narrowed, and the essay called that early. Review, cleanup, and advisory work grew instead, which the essay also called, without being able to name the cause. Automation created oversight, and almost nobody in 2018 priced that in.
So the honest reading for anyone working in QuickBooks is narrow and calm. The question was never whether accounting is dying. It is which tasks are dying, and the answer is the tasks a machine can finish alone. Scorekeeping is safe. Judgment is safer. The typing was the only part ever in real danger.