There are two conversations happening in finance right now about artificial intelligence, and both are wrong.

The first conversation says AI will replace finance teams. Controllers, FP&A analysts, and CFOs will be automated away. The spreadsheet is dead. The finance function as we know it is over.

The second conversation dismisses the whole thing. AI is a toy. It makes things up. You cannot trust it with numbers. The fundamentals of finance have not changed and never will.

Both conversations miss the point because both treat AI as something external to the finance function, something that either threatens it or does not apply to it. The reality is more useful and less dramatic than either position.

AI is an amplifier. It extends what a finance professional with strong fundamentals can do, faster and at greater scale than was previously possible. It does not replace the fundamentals. It makes them matter more.

The Iron Man Analogy

Tony Stark without the suit is already exceptional. He is an engineer, a strategist, a systems thinker. The suit does not replace those capabilities. It amplifies them. Flight, strength, processing power, real-time data integration — all of these extend what Stark can do, but none of them work without the person inside who knows what to do with them.

A pilot without the suit cannot fly the suit. The suit without the pilot cannot make decisions. The value is in the combination.

For finance, the dynamic is the same. A controller who understands financial statements, variance drivers, working capital mechanics, and business context can use AI to work faster, produce more analysis, and handle complexity that would have required a larger team. A controller who does not have those foundations cannot use AI to produce reliable output, because they cannot tell when the output is wrong.

And AI output is sometimes wrong. The relevant question is not whether AI makes errors, because it does. The relevant question is whether you have the technical grounding to catch them.

What AI Actually Does Well in Finance

The capabilities that are genuinely useful in a finance context right now are:

Processing and structuring unstructured data. Financial data rarely arrives clean. Bank statements, PDF invoices, email attachments, and poorly formatted exports all require significant manual processing before analysis can begin. AI tools can classify, extract, and structure this data at a speed that no manual process can match.

First-draft generation. Variance commentary, management report narrative, board pack text, and financial analysis summaries all require the same basic structure: what happened, why it happened, what it means. AI can produce the first draft of that structure in seconds. An experienced controller edits and corrects it in minutes. The alternative is writing from scratch, which takes much longer.

Pattern recognition in large datasets. Identifying outliers, spotting trends, flagging anomalies in transaction-level data — these are tasks that humans do poorly at scale and well in small volumes. AI inverts this. It handles scale effortlessly and surfaces the items that require human judgment.

Code and formula generation. Complex Excel formulas, SQL queries, Python scripts for data processing — AI can produce working code from a plain-language description of the problem. This removes a significant technical barrier for finance professionals who are not developers.

Scenario modelling acceleration. Building out alternative scenarios in a financial model requires applying logic consistently across many cells and variables. AI can translate a verbal description of a scenario into a structured set of model changes, dramatically reducing the time from question to answer.

What AI Does Not Do

AI does not replace financial judgment. Judgment is the ability to know which number matters, why a variance is significant, when a trend is real versus noise, and what a set of financial results means for the business. These capabilities are built from years of working in finance, making mistakes, understanding how businesses actually operate, and developing an intuition about what the numbers are telling you.

AI does not know your business. It has no context about your company’s cost structure, your customer concentration risk, your seasonal patterns, or the fact that the CHF 200k unfavorable variance in April is because of a one-time inventory writedown that does not repeat. You know that. AI needs you to tell it.

AI does not take responsibility. When the board asks why EBITDA missed by CHF 500k, the AI will not be in the room. The controller will. The ability to stand behind a set of numbers, explain the drivers, and propose corrective actions is irreducibly human.

AI also does not handle ambiguity well without guidance. Financial accounting is full of judgment calls: how to classify a cost, whether a provision is adequate, how to present a complex transaction. AI will produce an answer to these questions, sometimes a plausible one, but it does not have the professional training, the regulatory context, or the ethical accountability to make these calls reliably.

The Skills That Become More Valuable

If AI handles speed and scale, the skills that become more valuable are the ones AI cannot replicate.

Technical depth. The ability to understand what AI is producing and whether it is correct. A controller who understands IFRS, working capital mechanics, and financial model structure can verify AI output. A controller who cannot will not be able to distinguish good output from bad.

Business context. Understanding the business well enough to ask the right questions and interpret the answers. AI can produce analysis of anything you put in front of it. Knowing what to put in front of it, and knowing what the output means in context, requires someone who understands the business.

Communication. Translating financial complexity into clear, credible communication for non-finance stakeholders. This is a skill that AI can assist with but cannot replace, because it requires understanding the audience, the relationship, and the strategic context.

Judgment under uncertainty. Making a call when the data is incomplete, the assumptions are contested, and the stakes are real. This is where finance professionals earn their place at the table.

The Practical Implication

Finance professionals who ignore AI will find themselves doing manually what their peers accomplish in a fraction of the time. The gap will compound. In two years, the controller who has integrated AI into their workflow will produce two to three times the analytical output of the one who has not, at the same level of quality. That productivity difference will be visible to every CFO and finance director making hiring decisions.

But finance professionals who treat AI as a magic wand, generating output without checking it, building models without understanding them, and presenting analysis they cannot defend, will create expensive problems that discredit both themselves and the tools they are using.

The Iron Man suit makes Stark extraordinary. Without Stark, it is just a very expensive piece of hardware that falls out of the sky.

Learn the fundamentals. Then put the suit on.

Alessandro Ratzenberger is a fractional CFO and business controller based in Zurich, with 15 years of operational finance experience at Dufry Group and Bomi (UPS Group). Book a free 30-minute call or browse the finance templates.