When it comes to successfully implementing AI inside of their finance functions, today’s CFOs are weighing “two different aspects” against each other, said Vivek Saxena, SVP and service line lead for finance & accounting at SaaS provider Genpact.

Firstly, finance leaders are looking closely at costs, Saxena said in an interview — including both the costs of the technology itself and the expenses related to running a modern finance function. They are then weighing those costs against value, he said; both the value finance provides today, versus how the technology could impact the value that function could provide in the future.

For example, when it comes to something like improving forecasting, “that is not a cost game — that's a value from finance game,” Saxena said. “Once you get the right insights, how will you be able to monetize that or quantify that?”

Risk versus reward

The challenge of balancing short-term costs against long-term impacts is not a new one for finance chiefs. However, dong so with AI can be tricky for CFOs who need to keep pace with changing regulations, constantly evolving technologies, and macroeconomic factors such as pricing pressures to determine how implementing AI could impact their companies.

With the advent of new AI solutions, such as “agentic AI” tools, the focus for many finance chiefs is on risk management, Saxena said.

Presently, those looking to implement AI “will spend more time [figuring] out whether the adoption of these AI modules or agents will impact their existing control environment effectiveness,” he said. CFOs then need to determine what steps they should take to mitigate that risk.

Genpact offers several enterprise solutions supported by AI, and recently partnered with Databricks to build an “AI gigafactory” that will help financial service firms better move their AI products into production, according to reports. An 18-year veteran of the New York-based company, Saxena has served in his current role since September 2022, according to his LinkedIn profile. Before joining Genpact, he served as manager, CPM product management for Oracle.

Agentic AI has steadily filtered into the spotlight over the past year, with large-scale players such as Microsoft and Salesforce offering such agents to help financial leaders with key processes and workplace tasks, CFO Dive previously reported.

While AI agents could potentially help to improve profitability or reduce expenses by cutting down on the time it takes finance employees to finalize routine tasks — such as the financial close process — both industry leaders and regulators have expressed concerns about the technology’s use. The technology’s propensity to “hallucinate” or to provide erroneous information has raised flags, with questions also arising over its safety and security.

CFOs therefore need to pay careful attention to where they are looking to apply such tools and the reporting or regulatory standards when it comes to those processes, Saxena said. For example, “the dependency on traceability and auditability of AI models is much higher when it is…in the journey of preparation of financial statements versus when it comes to preparation of management reporting,” Saxena said.

The data mandate

The availability of AI agents may also lead to a shift in how CFOs think about their staffing needs, Saxena said. As such agents become more available, finance leaders may start to reconsider the ratio of employees versus automated tools, and where to apply each of them, he said.

“The challenge is, in many companies, the CEOs have given a mandate: adopt AI, deploy AI, and CFOs don't want to be behind,” Saxena said. AI and large language models have also steadily become more available for use in key financial processes such as accounts payable, or financial planning & analysis, Saxena said. At the same time, “training them [has] become comparatively cheaper,” he said.

That could lead to a potential future where companies are spending more on AI agents, shrinking the budget for hiring. As AI agents become more commonplace, certain finance jobs could ultimately disappear, Microsoft’s Georg Glantschnig, vice president of Dynamics 365 AI ERP told CFO Dive — though the shift will also make room for new jobs, he said.

There are still some roadblocks to implementing AI agents and related tools at scale, however. Costs are still high for certain key aspects of the technology, such as aggregating data.

Improving both the quality and flow of one’s data is a “multi-year journey” companies will need to undergo as the technology continues to evolve, Saxena said. To take full advantage of emerging technologies, CFOs need to have a clear understanding of both the reliability of their data as well as be able to trace where it’s coming from, he said.