A Heidrick & Struggles survey of C-suite executives found that most CFO respondents indicated their organizations have yet to use AI in crucial areas like auditing, accounting, financial reporting or FP&A. A Teradata survey published last week showed only 12% of very large companies have deployed AI company-wide.
However, these “studies” make a big assumption by hinting that finance teams are far behind on the GenAI adoption curve. They suggest AI products and services like large language models (LLMs) are ready to deploy, and all CFOs need to do is lay out the funds and plug in the tools.
In finance, for most midsize companies, that’s just not true. For them, job one is a solid foundational understanding of the current wave of AI technology and how it might change organizations and workflows. “Many CFOs are falling behind because they don’t know where to begin,” says Wayne Berson, chief executive of BDO USA.
There are plenty of descriptions of logical use cases for GenAI in finance—processing unstructured data to incorporate it into forecasts, extracting non-standard language from business agreements, serving as a helpdesk for suppliers who have questions about their invoices and using a tool like ChatGPT to gather and structure information for internal reports, to name just a few.
“Everybody is assessing what is real versus what are the possibilities,” says Vivek Saxena, a senior vice president and leader of finance and accounting services at Genpact. But “there’s a huge problem [not] having the right data. AI models depend on good data.” Successful deployment of GenAI also hinges on the ability to manage the potential cultural adaptations within the organization, including employees’ over-reliance on AI outputs or their reluctance to use the technology.
Building Awareness
One way to find out what GenAI can do in finance is to put it in the hands of employees. The primary goal of accounting firm Marcum’s GenAI initiative was to “become an AI culturally aware organization,” says Peter Scavuzzo, Marcum’s chief information and digital officer. That required getting the employee base comfortable enough to contribute to the firm’s “AI journey.”
Marcum put an internal AI tool into the hands of knowledge workers so it could crowd-source accounting and advisory use cases. “If somebody in the tax department goes in and tries to solve a problem using AI, we have a conversation with them and ask, What’s an interesting problem you tried to solve? What were you attempting to do? And what challenges did you have?” says Scavuzzo. A month ago, Marcum crossed 100,000 prompts, the goal being to publish a library of use cases.
“This is a cultural journey shift—a change management effort,” says Scavuzzo.
Reframing Expectations
If GenAI is genuinely going to make workers more productive, workers (and their CFOs) need to know “what AI can do, how to build the right prompt to get what they need and knowing when and where to use it, and how AI fits into their daily workflows,” said WalkMe CEO Dan Adika on his company’s May 22 earnings call.
Just as crucial as getting employees to experiment with the tools is ensuring they have realistic expectations of GenAI tools. Many business executives expect absolute answers and unassailable facts from AI output. When they don’t get them, they’re apt to think AI isn’t useful at all. But part of learning about AI is understanding what it can and cannot do. Scavuzzo compares using AI to an American football team advancing down the field. Using AI, you can’t go 100 yards, endzone to endzone, but it will get you in field-goal range nearly 100% of the time, he says. “Who wouldn’t want to advance 70 yards at the snap of a finger?”
Workers have to become aware of what GenAI is good at and not good at. “While the technology is absolutely moving to provide more precise and accurate answers, you can’t just blindly accept them,” says Scavuzzo.
Thinking About ROI
Running an internal GenAI tool can be expensive, and for now, costs vary widely. Genpact’s Saxena says the inputs to total cost may become somewhat more apparent in the next 12 to 18 months. For now, they may be hard to quantify. Not only does a company have to have the necessary tools and systems infrastructure, but some employees will spend time and energy on governance, including the drafting of usage guidelines and determining which use cases get approved.
In finance, “the CFO organization doesn’t want to be left behind when the CMO and the CIO are doing something in different parts of the businesses,” says Saxena. But that probably means going forward without an absolute answer to whether new ROI measures will be required to justify future spending, particularly when gauging AI contributions to worker productivity or even employee engagement.
Be Patient
“Without question, there are AI use cases very beneficial to finance organizations,” says Jotham Ty, founder and CEO of Gappify, a provider of accrual automation software for accounting teams. But the market is still sorting out which use cases to tackle and how they get rolled out.
“How this gets deployed and how audited, how it gets prioritized—I think that’s all still being deliberated,” says Ty.
Ty says Gappify has a good idea of how it will incorporate AI into its products, but it’s taking a cautious approach to the launch and promotions. “We don’t want to lean in until we have our messaging very clear and we have a good way of explaining to our prospects and our customers how they can adopt it and how to ensure auditability of the transactions,” says Ty.
“I know some have jumped the gun. We have some peers who are socializing features they will have for next year. That is not going to be us,” Ty says. “At the end of the day, as accountants, we value trust. And if we break that trust by promoting something we don’t have, it could be a deterrent.”
Ty feels the market will help with education, too. “One thing that will start emerging in the next couple of quarters is better framing of the different types of AI out there,” he says.
However, having the patience to wait for some of that doesn’t mean being passive. Scavuzzo says that armed with AI tools, diminutive, agile competitors with much less capital may use the technology to level the playing field in a market. Or an outside player may spot an opportunity to disrupt a sector because the sector’s entrenched competitors have been too slow to move forward with AI tools.
As a result, says Scavuzzo, CFOs need to think broadly about AI. The question is not just, “How can AI help me solve the problems I have today?” CFOs must also ask what it means for the organization’s customers, commercial markets and business partners.
A shorter version of this story appeared in the May 31 edition of the Finance & Accounting Technology Briefing.
