Casey Janick is by no means a Luddite. The founder of the CFO advisory practice CEL Capital and an operating partner with The CFO Alliance, Janick has helped clients avoid hours of work by automating daily processes like revenue reconciliation and champions the use of AI to speed budget analyses.
He also believes that AI’s penchant for speed makes it easier for some companies to go without a full-time CFO, strengthening the hand that fractional finance chiefs such as himself can offer clients.
But he has begun to notice a way that AI’s influence is creeping into finance conversations that is counterintuitively increasing finance workloads. What’s behind the shift? He believes investors, boards and lenders are increasingly prompting AI platforms to create questions about financial data that can send finance teams on a kind of wild data goose chase.
“There needs to be transparent conversation about the increase in workload that [loop] will create for the reporting functions because it’s essentially creating more rounds of review notes going back and forth,” he said in an interview. What’s increasingly missing, he said, are questions arrived at via natural human curiosity.
Number narratives
Janick spoke with CFO Dive in Oak Brook, Illinois, after a recent roundtable discussion held by the CFO Alliance professional networking group. During the meeting, CFO Alliance founder Nick Araco Jr. encouraged members to start asking individuals on their finance teams to defend their numbers and to work to develop their ability to explain why a given number matters.
Janick knows there are challenges to such skill development efforts. A CPA who previously worked in public accounting roles at both Plante Moran and KPMG, Janick acknowledged that structured training is less common in industry finance roles than it is at large accounting firms. On top of that, technology means there are often fewer people to teach in finance departments.
“There are less people able to ask and answer why than there used to be, and there are more tools to pretend you know why than there ever have been,” he said. Still, he maintains it’s important to teach financial professionals at all levels a central tenet: If you don’t have the answer to why the number is important, you do not have the answer.
Wrong directions
While some of the roiling debate over the risks of AI has focused on it developing “hallucinated” or incorrect data, Janick’s concern centers on what he sees as a rising false confidence in AI’s ability to develop questions and direction regardless of whether they are worth the time it takes to answer them.
At times, AI is “giving the wrong direction sometimes, focusing [companies] on the wrong things,” he said.
For example, he said companies have multiple sources of financial data, ranging from ERP systems to point of sale systems to payroll and benefit platforms. AI tools don’t always have the ability to make a determination as to whether the data that is siloed in those different areas is comparable. He has seen this mismatch leaving revenue officers and even CEOs who may have queried AI citing data that may be misleading, in part because there are inconsistent definitions of the measure that is being queried.
“There’s also raw data that isn’t financial, head count and certain employee benefits and things that come from HR ...[that] may not line up with what we are reporting and AI doesn’t care or have the ability to make that determination,” Janick said.
One of the key solutions to these AI shortcomings is better governance, he said, asserting that companies must work to establish a framework for using AI along with accountability. Janick recommends that companies set up an AI steering committee, ideally led by the chief operating officer or someone in IT. An important component of this is that the group includes all department heads, he said.
“I don’t see how a business is going to be able to function with data being combined if the people looking at it aren’t combined,” Janick said.
