According to Gartner, 90% of finance functions will deploy at least one AI-enabled technology solution by 2026, but less than 10% of functions will see headcount reductions.
The technological research and consulting firm says chief financial officers are already making changes to fully harness AI in finance. However, a sense of uncertainty, inflated expectations, and employee disengagement often quell success rates in AI usage.
CFOs who combine the strengths of people and machines increase their chances of AI success through a satisfied and engaged workforce.
"Despite AI’s ability to emulate human performance, algorithms cannot match the unique capabilities of people in areas that require creativity and complex problem solving," says Ash Mehta, senior director analyst in the Gartner Finance practice.
He says that by recognising the respective strengths of people and machines, finance leaders can build processes that boost the abilities of people and machines while mitigating their weaknesses. This requires a new kind of collaboration between people and machines that will improve business performance and employee satisfaction.
Mehta cites AI-driven machines as an example, noting that while these are very adept at automating simple decisions and processes by analysing large amounts of data quickly, they cannot work independently and may fail to form good conclusions when presented with unusual circumstances.
On the other hand, people use creativity and an innate understanding of human behaviour to conclude when presented with new and unfamiliar problems quickly but could not hope to outperform a machine when crunching numbers.
"To supercharge the abilities of both AI and people, they must learn to collaborate in a way that harnesses each other’s strengths,” says Mehta.
Gartner experts call this collaboration the human-machine learning loop, which promotes continuous process improvements that encourage finance staff and AI-driven machines to collaborate on processes while dividing labour according to the respective strengths of each.
While relying on each other for improvements, both parties can iteratively add greater value.
The human–machine learning loop starts with the creation of an algorithm, automated process, machine-driven task or autonomous workflow, respecting what machines are as good or better than people at performing. Then machines carry out these tasks, such as generating a revenue forecast, approving an expense report, or determining optimal payment terms for a specific customer.
Gartner adds that machines can also inform and advise. Such is the case when a forecasting algorithm suggests that a recent policy change will alter the sales outlook, or a machine-driven invoicing process suggests sending invoices on certain days to increase cash collections.
The labour performed by machines in this way then frees humans to take information, advice and recommendations from algorithms, using their creative and strategic strengths to solve complex problems by designing process improvements. Once new processes are in place, people trigger the next iteration of the loop by building new machines that execute the new processes and analyse the respective data.
