The short answer. No, not as a profession. The US Bureau of Labor Statistics still projects 5% growth for accountants and auditors from 2025 to 2035. But AI is already taking over routine work such as data entry and transaction coding, and the clerical roles built on that work are projected to shrink. Accounting is shifting toward review, analysis and advice.
On paper, accounting looks like the perfect job for AI: rules, numbers, documents and deadlines. That is exactly why the routine layer is going first.
The evidence so far points to a split. Routine tasks such as data entry and transaction coding are being automated, while checking, explaining and advising become a bigger share of the job. Here is what the data says, and how to stay on the right side of that split.
What does AI already do in accounting work?
AI already codes transactions, drafts memos, researches guidance and speeds up the month-end close. What it does not do is take responsibility. Both the PCAOB's outreach to audit firms and peer-reviewed field research describe people reviewing AI output and staying accountable for it. The table shows where the line sits today, task by task.
| What AI does today | What stays with you |
|---|---|
| Transaction coding and data entry. AI-enabled bookkeeping tools classify transactions. A peer-reviewed field study found adoption shifted accountants' effort "away from routine data entry toward business communication and quality assurance tasks" (Choi and Xie, Journal of Accounting Research) | Checking the uncertain items. The same study found accountants "selectively intervene when AI confidence scores are low" |
| Month-end close. AI use was associated with "more granular ledgers and faster month-end closing" in the same field data | Deciding what the numbers mean and signing off on them |
| Audit research and memos. Some audit firms let staff use generative AI for "initial drafts of memos and presentations" and to research accounting and auditing guidance (PCAOB staff) | Accountability: firms told the PCAOB that the person who uses an AI tool "is still responsible for the results and documentation of the work" |
| Audit procedures. Firms see future uses in summarizing accounting policy and legal documents and checking disclosures for completeness (PCAOB staff) | Professional judgment and supervision of the engagement |
| Advising clients and management. Automation frees time for it. BLS expects automation of routine tasks to "make accountants' advisory and analytical duties more prominent" (BLS) | The recommendation and the client relationship |
The same study carries a warning. In its framed field experiment, AI assistance improved classification accuracy on average, but "reliance on non-consensus AI recommendations can increase the risk of error." The authors conclude that AI works best as a tool that augments professional judgment rather than replacing it.
The accountants who gain the most from AI are the ones who know when not to trust it. In field data covering 79 small and medium-sized businesses, accountants stepped in when the AI signaled low confidence. Checking the uncertain cases, instead of approving everything or redoing everything, is a habit you can practice.
What does the job outlook for accountants say?
The US Bureau of Labor Statistics projects employment of accountants and auditors to grow 5% from 2025 to 2035, faster than the average for all occupations. It counts 1,595,200 jobs in 2025, a May 2025 median wage of $83,680, and about 115,300 openings a year. The occupation one rung down tells a different story: for bookkeeping, accounting, and auditing clerks, BLS says "software innovations have automated many of the tasks" they perform.
| Occupation | Change, 2025-35 | Median pay, May 2025 |
|---|---|---|
| Accountants and auditors | +5% | $83,680 |
| Bookkeeping, accounting, and auditing clerks | -6% | $50,670 |
BLS is explicit about AI. In its words: "Some routine accounting tasks may be automated as platforms such as cloud computing, artificial intelligence (AI), and blockchain become more widespread. Although these technologies will likely increase accountants' efficiency, this change is not expected to reduce overall demand."
Usage data points the same way. In Microsoft Research's study of 200,000 anonymized Copilot conversations, accountants and auditors score 0.19 on AI applicability in the published scores, roughly half the score of personal financial advisors (0.36) or customer service representatives (0.41). The authors note that relative comparisons are more meaningful than the absolute values. Accounting has plenty of AI-ready tasks, but a smaller share of the whole job than the most exposed roles.
The honest caveat: if you expected to learn the job through data entry and first-pass coding, plan for less of that work. Build review and analysis skills on purpose, and early.
Which skills keep accountants valuable?
The skills that keep accountants valuable are the ones AI cannot own: checking output, framing the problem, deciding what to delegate, protecting data and turning numbers into advice. None of them requires coding. All of them improve with short, regular practice on real work, which is how kju builds AI fluency.
- Checking AI output like an auditor. Treat every AI draft as unreviewed work from a new hire. Know how models make up plausible facts, and verify the figures and sources that matter before they leave your hands.
- Giving AI the context it needs. A generic prompt gets generic variance commentary. Audience, materiality, last month's commentary and a definition of good change the result. This is context engineering, and it is a learnable skill.
- Deciding what to delegate. High-volume, low-risk drafting is a good first job for AI. Anything that posts to the ledger or reaches a client without review is not.
- Protecting client data. Audit firms told PCAOB staff that data privacy and security remain a focus, and some described safeguards on what can be uploaded to AI tools. Know your firm's rules and why they exist.
- Turning numbers into advice. As routine work shrinks, explaining what the numbers mean for a decision is the part of the job BLS expects to grow.
For the wider set of skills every function needs, see AI skills every team needs.
Try three real kju questions for accountants
These come from kju's daily practice for finance and accounting roles, and kju users get eight like these each day, adapted to what they miss.
Question 1. You're automating the month-end close flow in Finance & Accounting at a consulting firm. Which step must keep a human in the loop first?
- A) Approving figures before they are booked or published
- B) Every step: automation should never gain autonomy
- C) None: partial automation isn't worth building
- D) Whichever step currently takes the longest
Answer: A. Keep humans where errors are costly or irreversible; expand autonomy as evidence accumulates. Source
Question 2. Which habits reduce the risk of acting on a wrong AI answer? Select all that apply.
- A) Check the sources that matter
- B) Treat outputs as drafts
- C) Use it where you can judge it
- D) Trust answers that sound confident
Answer: A, B and C. Verification habits beat trust: sources, a draft mindset, and staying inside your judgment zone. Source
Question 3. What's the root cause of LLM hallucinations?
- A) The model optimizes for plausible text, not for truth
- B) Bugs in the training code
- C) Insufficient compute at inference time
- D) Deliberate randomness added for creativity
Answer: A. Hallucination is a by-product of the objective: plausible continuation, not truth. Source
What should you do next?
If you're an accountant, practice a little every day on the tasks AI is already changing. Checking and framing are habits, and habits come from repetition. A kju session is eight questions and a real-work challenge a day, matched to your role: start a daily practice at kju.ai.
If you lead a finance team, the shift is a team change, not an individual one. kju trains whole finance teams with daily practice tailored to each person's role and level, and lets you see progress across the team. See kju for teams and how it works in financial services.
For the bigger picture across roles, read what jobs AI will replace, then compare the outlook for financial advisors and data analysts.
Frequently Asked Questions
- Will accountants be replaced by AI?
- Not as a profession, on current evidence. The US Bureau of Labor Statistics projects employment of accountants and auditors to grow 5% from 2025 to 2035, with about 115,300 openings a year. The routine layer is shrinking instead: BLS projects bookkeeping, accounting, and auditing clerks to decline 6% because software has automated many of their tasks.
- Will AI take over accounting?
- AI is taking over parts of it: transaction coding, first-draft memos, research on accounting guidance, and some close work. Audit firms told the PCAOB they expect generative AI to augment, not replace, people, and that the person using an AI tool stays responsible for the results. Review, judgment and advice remain human work.
- Is accounting still a good career with AI?
- BLS projects faster-than-average growth for accountants and auditors and reports a May 2025 median wage of $83,680. It also expects automation of routine tasks to make advisory and analytical duties more prominent. The safest path is to build those skills early, along with the habit of checking AI output.
- What AI skills should accountants learn first?
- Start with checking AI output, giving AI the context it needs, deciding what to delegate, and protecting client data. These skills apply whichever accounting software your firm uses, and they are the ones that keep you accountable for numbers you did not type yourself.