Will AI replace CPAs?

Not in the wholesale sense, but it will replace specific tasks that currently make up a large share of junior CPA work — data prep, reconciliation, document review. The role of the CPA shifts toward judgment-heavy work (interpretation, advisory, complex transactions) and toward AI orchestration (designing, validating, and overseeing AI-assisted workflows). The number of CPAs needed in 2030 is probably similar; the work mix is different.

Should accounting students change majors?

No. The accounting major teaches a way of thinking about money, controls, and risk that's not going away — and the demand for CPAs is structurally tight (declining test-sitter numbers vs steady demand). The skills that will matter most are the same ones that have always mattered (judgment, communication, ethics) plus AI fluency. Students who pair an accounting degree with strong AI skills will have an exceptional 2030 career.

Which accounting jobs are most at risk from AI?

The most-templatable, lowest-judgment work is most at risk: data entry, basic AP processing, single-entity reconciliations, mass-market tax-prep for simple returns. The least-at-risk work is judgment-heavy, regulated, or relationship-driven: audit conclusions, complex transaction structuring, advisory, controllership at small to mid-sized companies, fractional CFO work.

The CPA's AI Moment: What the 2024-2030 Transition Will Actually Look Like

In 2024 the CPA profession was talking about AI as a future inconvenience. By late 2025 it was being treated as a present reality. By 2030 the work mix of a typical accountant will look meaningfully different from today's. Here's an honest forecast of what changes, what doesn't, and how to position yourself.

What disappears (or shrinks dramatically)

Data entry and basic AP processing. Already 80% gone in firms that have invested in document-extraction AI. Will be 95% gone by 2028.
Bank and credit-card reconciliations for simple accounts. Largely AI-handled by 2027.
First-pass document review for audit and tax. AI sorts and triages; humans review the flagged items.
Variance commentary drafting. AI produces drafts; humans edit. Net time per report drops 60-80%.
Mass-market simple tax-prep. TurboTax-style tools become AI-assisted, handling more complex returns each year. The middle market for human preparers compresses.

What expands

Advisory work. When the data prep is automated, the bottleneck becomes interpreting results — which is judgment work, which is human work. Demand for fractional controllers, financial planning, and transaction advisory grows.
Audit and assurance over AI-driven processes. Auditors need to attest to whether AI-driven controls work. This is a new sub-specialty.
Complex transactions. M&A, restructurings, cross-border tax, derivatives — too much judgment, too much edge-case-handling for AI to own. Demand stays strong.
Regulatory and forensic work. SEC reporting, internal investigations, fraud examination — the human-required portion stays.
Controllership at small-to-mid-sized companies. Many companies still need a human controller; AI changes the day-to-day, not the role.

What gets created

AI workflow architects. Accountants who design and validate AI-assisted close and audit processes. Already a real role at Big Four firms.
AI auditors. Specialists in attesting to AI-driven financial controls.
Hybrid CPA-data engineers. Accountants who can build the integration layer between ERPs, AI tools, and reporting systems.
AI-augmented advisory. Partners who use AI to scale advisory work that previously required heavy associate-level support.

The math problem the profession is sitting on

The number of US candidates sitting for the CPA exam has declined every year for almost a decade. Demand for CPA-level work hasn't. Even with AI productivity gains, there's a structural shortage that doesn't close in this decade. This is the opposite of the "AI replaces accountants" narrative — it's a profession with too few people, getting AI tools that make each person more productive, in a market that still wants more output than the supply can deliver.

For an individual CPA in 2026: pricing power for senior people is increasing, not decreasing. The squeeze is at the bottom of the org chart, where the work being automated lived.

How to position yourself

If you're a student

Pair the accounting major with measurable AI fluency: take a Python class, build something with Claude Code, demonstrate AI-driven productivity in your first internship. The candidates who walk into 2027 hiring rounds with both will be unusually competitive.

If you're a junior associate

Become the team's AI-tools expert before your manager does. The visibility is enormous. Don't wait for formal training — build something on personal data and demo it.

If you're a senior

Move toward judgment-heavy work fast. The work that defined "senior" five years ago — managing teams of associates doing data prep — is shrinking. The new senior role is closer to "advisor with AI leverage."

If you're a partner

Don't bet your practice on the work mix that built it. Reinvest in advisory and judgment-heavy work. Reorganize your team structure for fewer junior bodies and more senior leverage. The firms that don't will see margin compression they can't recover from.

The bottom line

The 2024-2030 transition is real and it's big. But "AI replaces CPAs" is not the right framing. "AI rebalances the CPA work mix toward judgment and away from data prep" is closer. The profession ends up with similar headcount, more productive per person, and a higher-leverage work mix. The individuals who rode the transition deliberately come out ahead. The ones who didn't will be okay too — they'll just be doing what less-AI-fluent peers can't avoid doing, slower than they need to.

The thing not to do is wait for the dust to settle. By the time it does, the early-adopter advantage will be gone.