AI will make your best accountants more valuable, not obsolete

The process—data entry, reconciliations, report generation, first-pass close work—is ripe for automation. No argument there. But the discipline? Interpreting standards for a novel contract structure, assessing whether an instrument warrants liability or equity treatment, evaluating lease modifications, constructing a position you can defend under audit scrutiny—that's not a data problem. It's a judgment problem. And judgment is precisely where AI remains structurally limited.

CFOs who confuse the two will make expensive mistakes.

Judgment can't be automated

Accounting standards are built on professional judgment by design. The FASB and IASB write guidance that says "it depends on the facts and circumstances" because they can't anticipate every structure a creative finance team will invent. A technical accountant doesn't look up an answer. They construct an argument, weigh competing interpretations, and defend a position.

AI can synthesize existing positions. It cannot originate a defensible new one in a gray area.

And here's the deeper problem: the most consequential technical questions arise when something is novel—a new business model, a new instrument, a transaction structure no one has seen before. These are the moments that matter most to a company's financial reporting. They're also the moments where AI, trained on historical data, is least reliable. A model that can hallucinate a citation from ASC 805 is not a model you want making your consolidation decisions.

Accountability isn't optional

When the SEC sends a comment letter, someone has to own the judgment call and defend it. When an auditor pushes back on a revenue recognition position, someone has to sit across the table and make the case. AI can't sign a 10-K. Regulators and audit committees don't want a probability score—they want a person who can be held accountable.

The stakes make this non-negotiable. Getting a routine close wrong is costly. Getting a revenue recognition policy or a consolidation determination wrong is a restatement, an enforcement action, or worse. The risk is asymmetric: the upside of automating technical judgment is modest efficiency gains, and the downside is existential.

The real opportunity

AI will make strong technical accountants dramatically more productive. It will handle the process work that used to consume their time, freeing them to focus on the judgment calls that actually protect the company. The premium on deep technical expertise won't shrink—it will grow, because once AI handles the routine, judgment is all that's left.

The CFOs who thrive won't be the ones who cut their technical accounting function. They'll be the ones who redeployed it toward higher-leverage problems.

But picture the alternative: you've automated aggressively, your team is lean, and now you're sitting across from your auditor with a question about a complex transaction—and no one in the room can defend the position.

That's not an efficiency gain. That's a vacancy where expertise used to be.