Major companies are imposing stricter controls on employee use of artificial intelligence tools as spending on AI tokens rises faster than many businesses expected. Accenture, one of the world’s largest consulting firms, has begun urging staff not to use costly AI services for minor tasks such as turning PDF files into basic presentation slides. The move marks a notable shift for a company that had previously pushed employees to adopt AI tools more aggressively, even warning that failing to use them could hurt promotion prospects.
At Accenture, AI strategy chief Justice Kwak said during an internal meeting that growing AI expenses are beginning to weigh on the company’s financial structure. He said senior finance and technology leaders are questioning whether the money being spent is producing sufficient value. Large language models such as ChatGPT consume tokens with each request, and companies that once encouraged broad adoption are now reexamining whether routine or low-value uses justify the cost. Similar guardrails are appearing elsewhere: Match Group now gives employees AI budgets through department heads, requires justification to exceed limits, and restricts default access to its most expensive models.
Supporters of tighter controls say the measures are a necessary step in managing a fast-growing expense category. Finance executives at several companies are taking a larger role in deciding who gets access to AI tools, which vendors are approved, and whether AI investments are delivering measurable returns. At Elevance Health, the company routes employee queries to different AI models depending on the complexity of the request in an effort to contain costs. Some executives argue that, when used carefully, AI can still pay for itself. Elevance Health’s CFO Mark Kaye said AI automation has reduced administrative work tied to medical-chart reviews by about 40%, while Zocdoc finance chief Netta Samroengraja said companies should continue investing when results are clearly measurable.
The shift is also feeding broader caution across the technology sector. Analysts have described a recent decline in shares of AI-linked companies, particularly memory chip makers, as an "AI selloff" driven by investor focus on revenue and return on investment rather than novelty alone. Companies are also reassessing hiring, software duplication, and vendor selection as AI spending expands. Executives at firms such as Xero say finance, technology, and HR teams are now working more closely and more frequently to judge which tools are worth keeping. Taken together, the changes suggest the corporate AI boom is entering a more disciplined phase, with businesses no longer treating AI budgets as open-ended and increasingly reserving advanced tools for higher-value work.