The costs of AI tokens, or the amount of AI processing power companies purchase, has fallen under scrutiny in recent months as companies seek to keep budgets balanced amid continued economic pressures.

Those pricing models may be beneficial for vendors, because “the sky’s the limit in terms of how much you can charge,” Patrick Villanova, CFO of BlackLine said.

“But it’s very challenging for people like me, for finance professionals that are held accountable to a budget that need that predictability, that need that line of sight, and that has to see an ROI on every investment,” Villanova told CFO Dive in an interview.

Looking for outcomes

The expenses associated with token-based models have led to rising frustrations for company executives — especially finance leaders, who need to be able to keep an accurate and transparent account of spending, and who are increasingly being asked to show a return on investment from their AI spending.

The lack of predictability that comes along with token-based pricing can be a particular sticking point, Villanova said, citing conversations with other finance leaders.

“You want predictability. You want line of sight,” he said, both of himself and fellow finance chiefs. “They want to understand, what benefit are you bringing to our business, and link it to metrics that are important to their business.”

Employees can burn tokens by experimenting with new iterations of code or by trying to generate new images or copy, leading to a spike in cost. Rideshare firm Uber, for example, exhausted its full 2026 AI token budget in just a few months, before instituting a $1,500 spending cap per month, per tool in order to keep costs down, according to reports by Fortune and Bloomberg.

A 10-year veteran of the Woodland Hills, Calif.-based company, Villanova has served as BlackLine’s CFO since March 2025, according to his LinkedIn profile. Before taking the CFO chair, he served as its chief accounting office for six years, and joined BlackLine after a 16-year span at Big Four firm PricewaterhouseCoopers.

BlackLine, an agentic financial operations platform which offers products for account reconciliations, journal entries and compliance, among others, offers “outcome-based,” rather than token-based pricing, he said. Rather than clients purchasing a particular number of tokens, they instead purchase an “outcome,” such as a number of reconciliations they want automated, he said.

BlackLine itself still purchases the tokens, which means the business has to be “very, very cautious that whatever price we get over here, justifies the rate of token consumption over there,” he said.  

Villanova takes that same outcome-driven approach when evaluating internal solutions for use by BlackLine, he said.

“What I mean by that is, when a budget owner or another executive comes to me and says, ‘I need X dollars to buy this AI product,’ my first thing is I say, ‘Okay, what outcome do you desire?’” he said. “And I'm like, ‘let’s not even talk about the cost yet. What is your goal?’”

Prioritizing transparency  

When it comes to finance and accounting, the cost of AI has not necessarily meant that CFOs or finance teams are abandoning experimentation with the technology, but that’s particularly because for finance, “adopting AI is intentionally a little slower,” Vilanova said.

The finance industry is regulated, and needs to report its figures to agencies such as the Securities and Exchange Commission or Public Accounting Oversight Board, as well as internal auditors, and “what they need to see is, it's got to be a glass box, not a black box,” he said.

“You have to have absolute transparency of what the AI is doing,” Villanova said. “You have to have absolute transparency of the decisions it's making and why, if it's making its own decisions or exercising judgment, you have to have a very clear audit trail.”

With the need for both transparency and predictability growing, Villanova predicts outcome-based pricing will become more common in the future. User-based pricing is “going the way of the dinosaur,” Villanova said. “It already has been for several years, because the more efficient you become, the less licenses you need, and it's kind of a paradox where the better you use our product, the cheaper it gets because you need less and less people using it.”

That trend is only going to accelerate with the continued introduction of AI, he said.

“There's going to be less and less accountants, less and less finance professionals doing some of this rote work that we're automating with agents, so now we're saying, look, one flat fee per year,” he said.