CFOs learning to balance the need for financial accuracy with the need for AI innovation need to start with “ambition,” Basware CFO Martti Nurminen believes. It’s important for finance chiefs to devote time to study AI, speak with their teams and external parties, he said.
“It's a huge change in my own calendar,” Nurminen told CFO Dive in an interview. “If I go back a year and I look at time that I allocated to AI [then] as opposed to today,” it’s a fundamental shift, he said.
The financial goalkeeper
Setting and maintaining key financial guardrails has long been an essential part of Nurminen’s strategy as a finance chief — an approach he learned playing soccer in his native Finland, where he was a goalkeeper. There are key parallels between the roles of goalkeeper and CFO, he said.
“You need to look ahead. You need to kind of play out situations before the actual things happen,” Nurminen said. “On a great day, you can even save the team with a great save. On the other hand, if you screw up, it's immediately at the scoreboard.”
Nurminen has served in the top finance seat for the Finland-based Basware, which offers invoice automation software, since 2019, according to his LinkedIn profile. Previously he was CFO for Affecto, another Finnish firm, and held numerous roles during nine years at software provider IBM, including as its group controller and finance transformation manager.
When it comes to developing the necessary guardrails for AI, one of the fundamental areas that is often overlooked is the relationship between the typically deterministic model of finance — where finance leaders need to produce clear, rules-based, verifiable numbers — and the largely probabilistic model of AI, which provides likelihoods, he said.
“There's acts in finance that always need to be deterministic, well controlled, and then there are areas in finance where we need to apply probabilistic techniques and also rely on probabilistic AI,” Nurminen said.
“I believe finance leaders, myself included, need to spend more time in getting to clarity about how this division plays out in our processes and in our outcomes,” he said. CFOs need to examine each process and deliverable carefully to see where and if AI can be beneficial, he said. In making those determinations, accurate financial data remain the bare minimum, meaning CFOs need to think carefully about what processes and functions will truly benefit from probabilistic AI.
“We can't have a situation that we ask a question now, and then I ask the question five minutes later, and I get a different answer, because that's fundamental,” he said.
Driving a competitive advantage
As AI becomes more integrated into businesses, “the job of finance leaders increasingly will be to set those policies, as well as the rules and regulations,” Nurminen said. “In our case as an example, at the lowest level, what's an acceptable tolerance on an invoice level? What's an acceptable price discrepancy in matching?”
CFOs expect to take more ownership of AI governance over the next few years: By 2030, 56% of finance chiefs expect to hold more responsibility for designing both financial and ethical guardrails for AI, a recent survey by IBM’s Institute for Business Value found.
Seventy-nine percent of the 1,500 CFOs surveyed, meanwhile, said they co-design AI governance, decision rights and risk thresholds with their technology leaders, the survey found. Another 62% said they had assumed new enterprise technology or “AI strategy leadership responsibilities” in recent years.
In order to ensure they are able to design those systems effectively, making time to gain more AI skills will be critical.
“CFOs need to be proficient in the technology that is closest…to the heart of the competitive advantage of their company,” Nurminen said.
