AI has spread beyond accounting: the firm's fight for position toward 2027

It's tempting to read the latest AI statistics as a single story: the curve is going up, and steeper than before. That's true. But the most interesting information in the numbers isn't how many companies use AI. It's where in the business AI gets used, and by extension, who owns the conversation about AI with the accounting firm's client.

The numbers, briefly

We wrote last week about the gap SSB documented between large and small firms, 8 in 10 versus 4 in 10, a gap that keeps widening year over year (SSB). That article was about who uses AI. This one is about something else: what they use it for, and why that changes the firm's position.

What has changed isn't just how many, but what

In 2023, accounting and financial management was the leading use case for AI in Norwegian business. By 2026, organizing administrative processes and business management, along with marketing and sales, have also become among the most common use cases (SSB).

We read that shift as something other than plain growth on several fronts at once. Accounting wasn't the leading use case in 2023 because accounting is the most important function in a business. It was the leading use case because accounting was the easiest place to start: structured data, clear rules, measurable impact. Vouchers, coding, and reconciliation were AI's entry point into Norwegian business precisely because the accounting profession had already put the data in order.

SSB points to generative AI as the main driver of growth in 2026. And generative AI lowers the barrier somewhere entirely different: in the unstructured. The meeting notes, the proposal letter, the job ad, the routine description, the customer email. The skill required to get started shifted from "system" to "language," and that gave AI more natural entry points into the business than the finance function alone. The consequence for the accounting industry is precise: in 2023, much of the client's AI conversation happened on the firm's home turf. In 2026, it also happens in administration and in the sales department, where the firm isn't at the table. The firm doesn't have to have lost ground for the conversation to have moved. It has simply gained more home turf than the firm's own.

The impact zone is the firm's own client base

It's worth noting which industries are moving fastest. Other service industries, manufacturing/power/water and sewage, and construction increased their share of AI users by 22 percentage points from 2025 to 2026 (SSB). Several of these are typical accounting-firm clients, manufacturing, construction, and power in particular.

That means a contractor client or an industrial client in 2027 may well have more hands-on AI experience from their own administration than the firm has to offer in its advisory work. Not deeper subject-matter understanding, but more practical experience with what the technology actually does in a working day. That's an uncomfortable, but entirely concrete, asymmetry to bring into the next client meeting.

The blind spot: under ten employees doesn't exist in the statistics

Then comes the part of the data that rarely makes it into the summaries. SSB's survey covers companies with at least ten employees. That means the core segment of the typical Norwegian accounting firm, companies with one to nine employees, isn't measured at all.

As far as we can tell, there is no national picture of AI maturity among the smallest businesses. Not from SSB, not from the industry associations, and not from the software vendors, who see usage of their own system but not the client's overall technology use. The one party with structured, recurring, voucher-verified insight into what's actually happening inside a four-person business is the accounting firm. An AI tool shows up as a subscription cost in the books long before it shows up in a survey. The firm sees the purchase, the frequency, the amount, and over time, whether it's followed by a change in hours billed, staffing, or revenue. That's a data access most firms are sitting on without having priced it.

Why the gap between 8 in 10 and 4 in 10 is an industry matter

The gap between large and small is widening, not narrowing. Large firms have their own specialist teams, their own project leads, and their own budgets to interpret the technology. The small business has no AI function. It has a CEO, maybe an office manager, and an external accountant. So the gap in the SSB numbers is less a technology gap and more an interpretation gap. Access to generative AI is the same for both; what differs is the capacity to judge what's worth adopting, what it costs, and what it does to the numbers. That's exactly the capacity an external finance function sells, and exactly the same expertise we wrote about last week that can't be divided into fourteenths, but can be shared across an entire client portfolio.

What the 2027 plan is really about

This is where the distinction Capassa sees as decisive sits, in the budget work now underway for 2027. Further automating the internal voucher flow delivers margin. It doesn't deliver position. It's nearly invisible to the client, because the client never saw the voucher work to begin with, and because competitors are doing the same thing at the same pace. The position sits somewhere else. Once AI has gained more home turf in text, processes, and sales, it still lacks one element that generative technology doesn't produce on its own: verified numbers and financial consequence. The client can draft the proposal with AI, but not judge whether the margin holds. The client can generate the routine description, but not see what it does to the wage cost.

It's in financial management and advisory work that the firm holds a natural monopoly on connecting AI use to the bottom line. The accounting industry was AI's entry point into Norwegian business in 2023. The question for 2027 isn't whether the industry can automate more. It's whether it can become the place clients come to find out whether their AI investments are actually paying off.

The statistics say AI has spread beyond accounting. They say nothing about accounting being unable to keep up.