Getting employees ahead of the AI learning curve: Tipalti CFO

AI’s potential has continued to captivate executives, but it’s important for leaders to create spaces and opportunities for their employees to experiment, said Sarah Spoja, CFO of AP automation provider Tipalti. The company offers automated accounts payable and payments management software to clients.

“I do think there is a learning curve for that and leaders need to push and give their team time and ideas and get them to start using this stuff, even before maybe it’s…ready for prime time,” she said in an interview.

Creating an AI culture  

Spoja has served as finance chief for the Foster City, California-based company since 2018, according to her LinkedIn profile. She joined Tipalti after serving as a director for KKR Capstone and as a consultant for Bain & Co., coming to the CFO role from the strategy or consultant side rather than the traditional accounting route.

Those experiences color how she approaches growth, where she spends her time as a finance leader and how she thinks about hiring, she said. Last year, Tipalti reported a 41% jump year over year in its customer base, growing to more than 3,500 customers and processing over $50 billion in payments, according to a press release.

When it comes to examining how emerging technologies like AI will affect the financial sphere, it’s important to seek out individuals that have a curiosity about new technologies and a willingness to experiment.

“I’ve already had for many, many years, a bent towards people who are a bit more entrepreneurial, and a bit more technology-driven inside my organization,” she said. Especially because Tipalti is a company providing products for the office of the CFO, it’s important to have team members who can spend some of their time thinking of how to make the products better, she added.

However, while it’s critical to keep that entrepreneurial spirit in mind when recruiting, that doesn’t necessarily mean searching for an individual who can fulfill a single, AI-focused role in the team, she said. Rather, the shift needs to be holistic throughout the organization.

“This is not something where you get one expert in your finance team who’s an AI expert, and they’re going to just change the way everyone else works,” Spoja said. “Everyone really needs to adopt it to be successful for the future.”

For CFOs and other executives, it’s essential to create space for their employees to experiment and to figure out the best way to tap the technology for their products and company.

“You’ve got to figure out and prioritize — and for every business, it’ll be different — what is the biggest opportunity for this within my organization?” Spoja said.

Fixing payment pain points

Creating that space to explore for one’s team members is especially critical in spaces like payments, where AI has the potential to fix many of the pain points today’s companies deal with routinely.

Despite the steady digitization of consumer payments, business payments have remained stubbornly manual, swinging open an increasingly wide door for fraudsters. In its latest Payments Fraud and Control Survey Report, the Association for Financial Professionals found 80% of companies reported being the victims of payments fraud or attacks in 2023. Of those that were the victims of this type of fraud, 30% were unsuccessful in recovering their lost funds, the survey of over 500 treasury professionals found.

When it comes to bringing the technology into payments, therefore, it’s important to consider not only how AI could help with the transaction, “but also the workflow for the data,” Spoja said. “And how does all of that integrate back to these core systems?”

For example, late last year, Tipalti launched an “auto coding” feature which taps generative AI to analyze the contents of key documents such as purchase orders and invoices. Such features ease the need for employees to review each item on a purchase order line by line as well as improving accuracy and “ensuring that basically anything that’s payments-executed is done through the platform,” Spoja said.

“The user rights and hierarchy in the platform ensure that the right people are signing off on payments and there’s sort of a log of it,” she said. “So no one can go in and change the banking information without the user rights.”

While some of that involves the automation Tipalti has traditionally utilized inside of its platform, emerging natural language models help to code things more appropriately and to streamline the process, Spoja said.

“The payments execution piece can be kept with a very [small] number of people, with very clear user rights,” she said. “Versus if you do this manually, you might have a lot of people across all of your banking portals…and it’s harder to control and streamline.”