I've been a CPA for over a decade. In that time, I've watched accounting go from spreadsheets and paper trial balances to cloud software to AI tools that can categorize transactions, flag anomalies, and generate financial reports faster than most senior staff.
Every few months someone asks me some version of the same question: "Should I be worried about AI taking your job?"
My honest answer: not the job I do today. But the one I did in 2012? That one's already gone.
What AI Is Actually Doing in Accounting Right Now
The capabilities are real, and they're growing fast. Modern AI tools integrated into platforms like QuickBooks, Xero, and standalone solutions can already handle a meaningful portion of what used to occupy hours of staff time:
- Transaction categorization. AI models trained on millions of bookkeeping records can classify most transactions correctly on the first pass, and they improve over time as they learn your chart of accounts.
- Anomaly detection. Tools can scan a general ledger and flag entries that look unusual: duplicate vendor payments, amounts that fall outside historical norms, entries posted to unexpected accounts.
- Bank reconciliation. Matching bank feed transactions to ledger entries is increasingly automated. What used to take a bookkeeper several hours per month now takes minutes of review.
- Financial report generation. Pulling together income statements, balance sheets, and cash flow summaries with period-over-period comparisons is largely push-button today.
- Document extraction. Receipt capture, invoice processing, and data extraction from PDFs have become remarkably accurate. The manual data entry that once defined entry-level accounting work is quietly disappearing.
Early in my career, a significant portion of billable time went toward exactly this kind of work: tedious, but necessary. Data entry. Account reconciliations. Pulling together trial balances. Formatting reports. Clients paid for it because it had to be done, and it required someone who knew what they were doing. AI is taking over those tasks. I won't pretend that's not a meaningful change for entry-level accounting roles, because it is.
Where AI Still Falls Short: The Judgment Gap
Here's what AI doesn't do well, at least not yet: the part that matters most to clients.
Judgment. Context. The ability to look at a set of numbers and tell you what they mean for your specific business. Not just what they say on the page.
When a client's gross margin drops 8 points in a quarter, AI can flag the variance. What it can't do is know that the client just switched suppliers due to a six-week delivery backlog, and that the margin will recover once the new contract terms kick in. That conversation happens on a call, drawing on months of relationship context. That's what professional advice actually is.
This isn't a gap that better training data is going to close overnight. The nuance involved in advising a specific business — with its specific history, its owner's specific risk tolerance, its specific mix of customers and vendors. That's fundamentally different from pattern recognition across large datasets. AI is exceptionally good at the latter. It's still limited on the former.
There's also the question of what happens when AI gets it wrong. And it does. A tool that confidently miscategorizes a fixed asset purchase as an operating expense creates a tax issue that takes longer to unwind than it would have taken to handle correctly in the first place. Someone has to catch that. Review the outputs. Exercise judgment about what's accurate. That's still the job.
AI can flag the variance. It can't tell you what it means for your specific business, or what to do about it.
What This Means If You're a Small Business Owner
Here's the practical takeaway: AI software doesn't replace professional guidance. What it does is free up your accountant's time so less of it goes toward mechanical tasks and more goes toward your actual situation. That's a trade worth making, but it works best when there's a professional in the loop.
An AI tool left running without regular human review is a liability, not an asset. Miscategorizations accumulate. Errors compound. And because everything looks clean in the dashboard, there's often no visible signal that something is off until you're sitting across from your CPA in April trying to unwind 12 months of bookkeeping problems.
What small business owners often don't realize is that the right advisor isn't competing with AI tools — they're using them. The question isn't "should I use AI software or hire a CPA?" It's whether the CPA you're working with is leveraging AI effectively to give you more value for your dollar, or whether they're still billing you for the manual work AI should be doing.
The firms and advisors who've adapted are spending less time on compliance busywork and more time on strategic questions: Is your entity structure optimized for where your revenue is today? Are you positioned correctly going into next year? Are there deductions or credits you're missing because no one's looked closely enough? That's the conversation that moves the needle.
What This Means If You're a CPA or Accountant
The profession is being pushed upmarket. If your value proposition is still centered on compliance tasks and data entry — tax prep, reconciliations, report formatting — it's worth thinking seriously about where you want to be in five years.
That's not a criticism. A lot of practitioners built good careers on exactly that work, and there's still demand for it. But the margin is compressing, and clients who understand what they're paying for are increasingly drawn to advisors who offer something more.
The advisory side of this work is harder to automate. Helping a client decide whether to elect S-Corp status involves tax law, California-specific costs, payroll infrastructure questions, timing rules, and an honest conversation about the client's operational reality. Helping a business owner interpret a difficult cash flow month requires knowing what's normal for that business, what's seasonal, what's structural, and what's worth addressing now versus watching. That kind of thinking isn't going to be replicated by a generalized AI model anytime soon.
The practitioners who will thrive in the next decade are the ones who recognize what AI is good at and use it, while doubling down on the judgment-heavy work that clients genuinely need.
The Real Value of a Good Accountant in the AI Era
I founded ADL Business Consulting partly because I believe small businesses deserve this kind of thinking year-round, not just at tax time. AI is part of how I'm able to make that work without charging enterprise rates. The automation handles the mechanical layer, which frees me to focus on the decisions that actually affect outcomes.
The best version of this relationship isn't human versus AI. It's a professional using AI to be more useful to you, faster, at a cost that makes sense for your business. The transactions get categorized. The reports get generated. And then someone who knows your business looks at what it all means and helps you figure out what to do about it.
That's not a job AI is going to take. It's the job AI is making more important.