Does enterprise artificial intelligence (AI) save teams money, or cost companies cash? It’s a question with a potentially category-defining answer, and it’s one that depends on whom you ask.
News broke this week, for example, that Apple is reportedly putting its employees’ AI usage on a tighter leash. The bigger question for corporate finance departments, though, is how limiting AI spending can save them real money without sacrificing growth.
The Cupertino technology and hardware giant is cutting the AI budget for certain internal teams by as much as a third, and other tech sector companies have also set limits on employees’ AI spending. Tesla has a $200 weekly limit, Uber has a $1,500 monthly limit and Meta has encouraged employees to use the company’s own tools, as PYMNTS covered.
The restrictions reflect a new reality for enterprise AI as its solutions scale and the models mature. Every additional prompt, automated workflow and software development task can generate another charge. Unlike traditional software licenses, usage-based AI costs can rise as employees become more dependent on the technology.
That creates an uncomfortable equation for CFOs. Cutting AI consumption can improve a technology budget while increasing labor costs, slowing development or reducing the quality of financial decisions.
PYMNTS Intelligence has been targeting that dilemma with months of original research. Findings from 2026’s data studies suggest the real challenge is no longer how much AI enterprises can afford but determining which AI expenses have become integral to the business.
See also: The Finance Stack’s Great Unbundling Has CFOs Asking What They Need to Own
Companies Have an AI Deployment Problem, Not an AI Access Problem
Enterprise AI adoption has moved faster than enterprise AI integration.
The PYMNTS Intelligence report “AI at Work: Why Deeper Enterprise Use Produces Stronger Results” found that among U.S. enterprises with at least $1 billion in annual revenue, 62% said they spend more than $10 million on new forms of AI in the previous 12 months, with the median falling between $10 million and $25 million.
The employees running up the largest AI bills may also be generating the greatest productivity gains. That’s why, for finance teams, AI utilization is becoming a misleading metric. The more consequential measure is how much it costs to complete a workflow, including the labor, technology and oversight required.
Research from PYMNTS Intelligence’s April installment of “The Enterprise AI Benchmark Report,” showed that 71% of executives at companies with at least $1 billion in annual revenue believe that organizational readiness is the chief limitation on AI performance. Only 11% said they think AI technology itself is the primary barrier.
Still, CFOs are realizing that the AI applications producing measurable efficiencies today may still be years away from recovering their development, implementation and integration costs.
PYMNTS Intelligence research published in August found that 90% to 100% of surveyed firms across healthcare, financial services and media rated their recent AI returns positively, but only 5% to 10% reported achieving full payback. Many expect a return on their investments to take five or six years.
Read more: The 7 AI Terms Every CFO Needs to Understand
The Highest AI Bills Could Be the Best Investments
A flat spending cap risks penalizing the intensity of AI usage rather than its inefficiency. That creates a paradox for usage-based cost controls, particularly across finance. The May installment of “The Enterprise AI Benchmark Report” from PYMNTS Intelligence found that banks, insurers and other financial companies have reached high AI adoption in 27 of 75 tasks examined.
Apple’s reported policy allows exceptions for employees requiring higher allowances. But for enterprises adopting similar restrictions, the economics will depend on whether approval processes can distinguish high-value usage quickly enough to avoid disrupting work.
That’s supported by PYMNTS Intelligence’s September findings in the report, “AI at Work: Why Deeper Enterprise Use Produces Stronger Returns,” which suggest that deeper integration, rather than deployment across more functions, is associated with stronger financial returns.
Related: Good CFOs Automate but Great CFOs Anticipate
AI’s Financial Value Is Moving Outside the Technology Budget
Not every AI return appears as a reduction in headcount or operating expenses. PYMNTS Intelligence research published in October found that 64% of surveyed finance teams use AI-generated insights or recommendations in cash-flow decisions, compared with 44% a year earlier.
Meanwhile, 28% of CFOs identified those insights as their most impactful cash-flow action, up from 13% in 2025.
An AI tool could therefore be expensive for the department paying for it and valuable to the enterprise funding it.
That’s the foundation-level reason that PYMNTS Intelligence data shows controlling consumption addresses only part of the enterprise AI problem. A business that caps AI spending but leaves inefficient workflows intact may save on computing costs while preserving the underlying expense. And a business that removes spending limits without changing those workflows could end up paying for AI without capturing its potential returns.
“The Agentic Edge Index: How Leading Companies Put AI to Work,” a PYMNTS Intelligence report from the Enterprise AI Benchmark Series, explores how large U.S. companies are moving from AI assistance to AI action and finds that 84% expect business growth without a matching increase in staff over the next three years. Meanwhile, 75% expect changes in employee roles and required skills.
For all PYMNTS B2B and digital transformation coverage, subscribe to the daily B2B and Digital Transformation Newsletters.
At PYMNTS Intelligence, we work with businesses to uncover insights that fuel intelligent, data-driven discussions on changing customer expectations, a more connected economy and the strategic shifts necessary to achieve outcomes. With rigorous research methodologies and unwavering commitment to objective quality, we offer trusted data to grow your business. As our partner, you’ll have access to our diverse team of PhDs, researchers, data analysts, number crunchers, subject matter veterans and editorial experts.
