Companies are reportedly turning to a variety of tactics to manage growing AI spending.
As the Wall Street Journal (WSJ) reported on Tuesday (June 30), artificial intelligence (AI) is increasingly being charged on a per-use basis, and companies are facing price fluctuations in the tokens that are a key component of AI computing.
“With AI, you’re putting the credit card in the hands of the end user, and if you don’t have control over it or the end user isn’t well-educated, they’re going to run up the tab,” Chris Reed, senior director of IT finance at online travel company Priceline, told the Journal.
According to the report, technology leaders at some companies are beginning to adopt cost management strategies developed during the rise of cloud computing.
Kathy Kay, chief information officer at Principal Financial Group, said financial services companies are “adopting governance and optimization practices similar to what enterprises have done in the cloud to manage the costs of scaling.”
For example, principals are focused on using the right AI models for the right tasks, so “higher utilization doesn’t necessarily mean higher costs,” Kay said.
“Given that prices and features are rapidly evolving, we are designing for flexibility so that we can adapt over time and continue to efficiently deploy AI,” she added.
Ravi Soin, Smartsheet’s CIO and chief information security officer, said the company’s FinOps (a combination of finance, engineering, and product) team is responsible for tracking overall AI spending. The software company created automatic alerts to notify employees that they are nearing their token limit.
“We have user dashboards available across the company, by department, and by manager, so we can see cost frequency and costs in real time, so there are no surprises at the end of the month,” he said.
Meanwhile, other companies are also benefiting from this trend, PYMNTS wrote earlier this month. For example, financial operations platform Ramp raised $750 million at a $44 billion valuation, nearly tripling in value in one year.
“The company believes that AI consumption, which is billed by tokens and fluctuates by prompt or agent action, has become a cost category that most corporate finance teams cannot track, allocate, or control,” the report said.
“The problems that Ramp targets did not exist at scale two years ago,” PYMNTS added. “AI providers initially priced access on flat-rate subscription terms. As the agent model moved to coding, customer service, research, and procurement, usage-based billing became the norm. Every step an agent takes runs a meter.”
