Will AI replace accountants? What changes, and where the work goes next

What AI already does in accounting

Automation in accounting is already happening inside the software most firms are using every day. AI tools now handle a large share of the routine production work:

• Data entry and transaction categorization. Bank feeds and AI categorization handle most of what used to be manual coding.
• Accounts payable. Bills are captured, coded and routed for approval with little human touch.
• Reconciliations. Matching transactions to statements is increasingly automatic, with a person reviewing the exceptions.
• First-draft reports. Standard financial statements, variance summaries and month-end reports can be drafted in seconds.

The early research on what this means for accountants is encouraging. A 2025 study from Stanford Graduate School of Business, drawing on survey responses from 277 accountants and task-level data from 79 small and mid-sized firms, found that accountants who use AI finalize monthly statements 7.5 days faster and spend 8.5% less time on routine, back-office processing. The researchers found that those accountants redirected the time toward client communication, quality review and client-facing advisory work.

That last finding matters a ton. AI gives accountants time back. What they do with that time decides whether they will be able to earn more.

What AI can't do

AI is very good at producing an answer - and confidently. It isn't accountable for that answer, it doesn't know the business owner, and it can't sit across the table and help them make a hard decision. Those gaps are where an accountant can continue to add the most value.

Being curious. AI will produce those confident answers without asking a single clarifying question. A curious accountant can ask good questions of their client to understand the decisions that drive the numbers before making an assessment or recommendation.

Judgment calls. AI can tell an owner that gross margin dropped three points last quarter. It cannot decide whether that's a pricing problem, a supplier problem or a one-time job that ran over, because that takes context about the business that isn't found in the books.

Accountability. When the reports are sent to a lender, an investor or the IRS, someone has to stand behind it. Clients pay for that signature and that responsibility, and AI can't take it on.

Explaining the numbers to an owner. Most small business owners don't think in financial statements. A good accountant translates the numbers into what they mean for payroll next month, the line of credit or the owner's own paycheck.

Deciding what to do about the numbers. This is the biggest gap. Imagine the client is a contractor whose cash is tight every spring. AI can produce a cash flow report that shows the dip. The accountant who knows the client can see that the dip comes from slow collections on two large jobs, recommend new payment terms on the next bid, and check in a month later to see if it worked. The report is the starting point. The advice and the follow-through are exactly what the client pays for.

Will AI take over accounting jobs? Who's exposed and who isn't

The truth is that it depends on the work you do today. Here's what the U.S. Bureau of Labor Statistics projects for 2025 to 2035.

Will AI replace bookkeepers?

Bookkeepers are the most exposed, and we should acknowledge and face that reality. The BLS projects employment of bookkeeping, accounting and auditing clerks to decline 6% from 2025 to 2035, a loss of about 85,600 jobs. The work being automated, like data entry, categorization and reconciliations, is the work at the center of most bookkeeping roles.

Bookkeepers also have an advantage that can be easy to overlook. You already know your clients' numbers better than anyone, and your clients already trust you. That puts you in a strong position to add a monthly advisory service, where you help those same owners understand what's driving their cash flow and decide what to do next. That's the work AI can't do. Our bookkeeper to CFO Advisor page explains how that move works.

Will AI replace CPAs and staff accountants?

Accountants and auditors are in a stronger position. The BLS projects employment to grow 5% from 2025 to 2035, faster than the average for all occupations, with about 115,300 openings per year. The BLS also notes that the automation of routine tasks like data entry will make accountants' advisory and analytical duties more prominent.

The pressure here shows up in pricing. When accuracy and timeliness are now table stakes, they stop being what sets a firm apart. Compliance work is becoming the baseline, and the fees for it are under pressure. The CPAs and accountants with the most room to grow are the ones adding work and fees that AI can't replace. Our CPA to CFO Advisor page explains how that move works.

Advisors and fractional CFOs

Advisors are the least exposed, because their whole service is the judgment and relationship work described above. AI makes good advisors faster. It does the analysis in minutes, which leaves more time for the conversations that actually help the client. Our page for fractional CFOs explains how our system fits into the practice you already run.

What we hear from accountants and bookkeepers

When accountants and bookkeepers come to The CFO Project, AI is one of the first things they ask about. The questions are usually some version of these:

The worry underneath is all the same. People can see the routine work going away, and they don't have a clear picture of what that means for them and the kind of work they can step into. The rest of this guide is about that picture.

Where the work is going: advisory

Every serious look at AI in accounting lands in the same place: the job is shifting toward advisory work. Most of those articles stop there. Here's what advisory work actually is, and what it pays.

What CFO Advisory is. CFO Advisory is a monthly service where an accountant or bookkeeper acts as a trusted guide for a small business owner. Each month you review the client's numbers, show them what's driving their cash flow, and agree with them about what to focus on and what to do next. You aren't the CFO of a Fortune 500 company. You're the CFO for a business doing $500,000 to $5 million a year and who they turn to when they want to understand their numbers and make better decisions. Our guide to what a CFO Advisor does goes deeper.

What the market shows. The AICPA and CPA.com 2024 CAS Benchmark Survey of 206 client advisory practices found a median growth rate of 17%, and median CAS (Client Accounting or Client Advisory Services) revenue up 61% over the 2022 survey. Only 10% of those practices still bill hourly as their main pricing method. The survey is self-selected, reflecting firms already investing in advisory, and the direction is clear.

What it pays. CFO Advisors in our program charge an average of $1,978 per client per month, and the service takes about 2 to 4 hours per month, per client. Five clients at $2,000 a month is $120,000 a year in recurring revenue.

How to move toward advisory work

You don't need a new degree or a new credential to start. You need a plan, a system and a few clients who already trust you.

Learning AI tools vs. learning to advise

A growing number of courses promise that if you learn the right AI tools or prompts, you can offer CFO-level services. AI tools are worth learning, and you should use them. But a prompt course simply doesn't solve the hard part.

AI tools are cheap and available to everyone, including your clients. Used without a system, they fall back on the average of everything they've read on the internet. You ask a real question about a real client, and the answer comes back confident, polished and too generic to use. A tool becomes useful when it sits on a system that tells it what to look at and what the client needs.

The scarce skill is turning analysis into advice a client will pay for every month. That means knowing which numbers matter for this business, having a repeatable process to review them, and being able to lead the conversation about what to do next. That skill comes from a proven system, practice with real clients and people who have done it before. It doesn't come from a better prompt.

That's why we built The CFO Project's system the way we did. It gives AI the structure it needs to do useful work, and it gives you a repeatable way to turn that work into advice your clients pay for every month.

How to use AI the right way in client work

If you're going to use AI with clients, a few rules keep the work useful and keep you protected.