From rear-view mirror to windscreen: What an AI-first CFO needs

In June, Boston Consulting Group published the article From Hindsight to Foresight, on how CEOs should give the CFO a new mandate in an AI-first world. The starting point is simple: Today, most finance teams spend roughly 90 percent of their time producing numbers and 10 percent advising the business. In an AI-first finance function, BCG writes, that flips. Collecting, reconciling and reporting numbers is largely automated, so the CFO and the team can spend most of their time interpreting results, giving advice and influencing the decisions that determine future performance.

BCG is writing for CEOs of large companies, and the results they cite, such as forecasting models that are at least 50 percent better, 90 percent automated reporting and more than 30 percent freed-up capacity in the finance team, come from companies at the front, not from the average. But we recognise the direction far beyond the large corporations.

The rear-view mirror takes most of the time

We often use a car as an image of financial management. The client sits in the driver's seat, and the finance function or the accountant sits in the passenger seat as navigator, not in the back with the receipts. The rear-view mirror shows what has happened, and that's necessary, but you don't get where you're going by looking in it. The 90-to-10 split describes a finance function that spends most of its time on the rear-view mirror.

The accounting industry has a similar shape. In the survey of member firms by Regnskap Norge, the Norwegian association for accounting firms, 91 percent of revenue in 2025 came from accounting and bookkeeping, year-end accounts and tax returns, and payroll and HR, while advisory accounted for 4.7 percent. These are two different measurements of two different kinds of business, time in a finance department and revenue in an accounting firm, so the figures can't be equated. But the pattern is the same: Producing numbers takes most of the capacity, and advisory gets whatever is left over.

A heads-up display, not a dashboard

BCG describes what it takes to shift the gaze forward: a "heads-up display" for the business. It isn't a traditional dashboard of backward-looking results, but a live overview of what drives the finances, such as sales activity, pricing, capacity and cost trends, linked directly to revenue, margin and earnings. Add real-time scenario analysis, and the CFO, the CEO and the board can see where performance is heading and which levers can still be pulled while there's time. BCG calls it a finance function with no surprises.

In our car image, this is the difference between looking in the rear-view mirror and having speed, fuel and the road ahead right in front of you on the windscreen. The result tells you how the engine is running, liquidity how much fuel is left, and the forecast where you're heading. If you only see the result afterwards, you discover you're out of fuel when the car has already stopped.

AI should sit on top of the general ledger

An important detail in BCG's article is what AI should not do. AI is still not 100 percent accurate, and when the numbers have to withstand audits and regulatory scrutiny, nearly right isn't good enough. That's why AI shouldn't replace the general ledger or the core systems that produce the numbers. It should sit on top of them and give the finance team deeper insight into why the numbers look the way they do, how they might change under different scenarios and what management should do next. That's also how we think about Capa: The analyses have to be based on the numbers actually in the accounts, and they have to be verifiable. BCG adds that prediction doesn't remove the need for human judgement, but demands more of it. Someone has to decide which signals matter, and when a person should override the model.

What we see

CFOs are an important customer group for us. There, it isn't about replacing the finance function, but about making it more efficient, so the CFO has more time to support the CEO. That's exactly what BCG describes as the goal: moving time from producing numbers to interpreting them. Magnhild Agerup-Faxvaag, CFO of ZTL Payment Solutions, puts it like this: "When the manual work disappears, I can spend my time on what actually creates growth." And she adds: "As CFO, I should be ahead of developments, not reacting to them."

The windscreen also shifts the conversation in the boardroom. "We use Capassa directly in our board meetings. Then we discuss decisions, not the underlying numbers," says Håkon Volden, CEO of Aarja Health. For him, it's about pace: "I don't have time to gather numbers. I need to understand them, quickly."

In many smaller businesses there is no CFO. Then it's the accountant who in practice sits in the passenger seat, and anyone who wants to help the client look ahead needs the same tools: a live picture of what drives the finances, early signals and analyses that can be verified.

The hardest part isn't the technology

BCG reminds us that the biggest obstacle for CFOs is probably human rather than technical. In their 10-20-70 principle, 10 percent is algorithms, 20 percent is technology and data, and 70 percent is people, processes and the work of changing how decisions are made. That matches our experience from building Capassa: Technology is rarely the constraint, changing the way people work is. The question for the CFO, and for the accountant, isn't whether it's possible to get a windscreen, but what you want to spend your time on when the rear-view mirror demands less of it.

A finance function with no surprises needs someone watching the road ahead, not just the road behind.

Sources: Boston Consulting Group, From Hindsight to Foresight: The CEO Mandate for an AI-First Chief Financial Officer (June 2026), and Regnskap Norge, survey of revenue by service area (2025, self-reported, split in kroner). The quotes from Magnhild Agerup-Faxvaag and Håkon Volden are translated from Norwegian. BCG's results come from leading companies and can't be generalised to all businesses.