AI Won’t Replace Accountants, But It Will Change the Job

New technology usually earns suspicion before it earns trust, and AI is no exception: for many people the first question about it is whose job it takes. In accounting that question starts from the wrong place. The profession’s problem is not too many people for the work. It is too few.

In a global survey of about 500 accounting leaders, 73% said their firms were turning away clients because they did not have the staff to serve them. When demand already outruns the people available, the most useful thing AI can do is give the accountants a firm already has more capacity.

The gap is demographic, and it was opening long before anyone put a language model near a ledger. Experienced accountants are retiring or selling their practices, and for years many younger people chose other fields that paid better and asked for fewer brutal seasons. The Bureau of Labor Statistics projects about 115,300 openings for accountants and auditors every year through 2035, many of them left behind by retirements and career changes.

The pipeline is recovering: US accounting undergraduate enrollment rose 8.9%, the third annual increase in a row. But a student who enrolls today needs four years of study, then experience, then a credential such as the CPA or CMA before they can carry the load of the person retiring. And accounting expertise compounds with years in the field, so what leaves with a retiring partner is not one headcount but decades of judgment that a new hire cannot replace on day one. For a while yet, experience will walk out faster than it can be rebuilt.

Accounting has always moved forward by getting more efficient, not by keeping work manual to protect the jobs attached to it. Given how short the supply of accountants is, the realistic effect of AI is to fill the gap rather than widen unemployment. Productivity lets the same team serve more clients, takes some of the pressure off people who are close to burning out, and lets firms hire because they choose to rather than because they have to.

The early field evidence points the same way. A Stanford Graduate School of Business study of 79 small and midsize firms found that those adopting AI supported 55% more client work each week, moved about 8.5% of accountants’ time from routine data entry to client communication and quality assurance, and closed the monthly books 7.5 days faster. The useful measure of success is capacity per accountant — clients served, speed of close — not how many people a firm managed to let go.

With AI as a copilot, the repetitive layer moves to software: data entry, reconciliations, numbers carried from one spreadsheet to the next. What remains for the accountant is reviewing that output, making the calls it cannot, explaining the result to the client and owning it. Critical thinking stops being a nice-to-have and becomes the job.

Large firms are already acting on that. EY announced a $100 million investment in rewarding its people, partly around judgment, adaptability and business acumen. In KPMG’s 2026 survey of summer interns, 76% said a successful career will take both strong human skills and the ability to direct AI well, with critical thinking and problem-solving ranked highest. For anyone building a career in the field, technical proficiency is still necessary; review, judgment and communication are what will set people apart.

None of this means nothing changes. Some roles will look very different, firms will make different decisions about hiring and margins, and some tasks will need far fewer people than they do today. But in a profession with persistent unmet demand, AI is far more likely to expand what existing teams can deliver than to shrink them. Entry-level work shifts from producing the first draft to reviewing what was produced automatically.

The harder problem is training. The traditional model had juniors spend years on repetitive work before they were trusted with judgment, and AI is removing much of that base. If the profession does not deliberately build judgment earlier, it risks a generation of accountants who never got the repetitions — and a career path that looks less inviting to the young people it needs. I have written separately about what juniors can do about that, and about how AI is likely to reach the firms that need the capacity most.