Companies are putting budgets and limits on workplace AI

Goldman Sachs chairman and CEO David Solomon said the bank is expanding the AI tools available to its 45,000 employees, but not everyone needs daily access to the most advanced models. “We need frontier models for certain things, particularly cyber,” he told Bloomberg TV, adding that less advanced tools can still help employees work more productively.

Cisco took a staged approach, chief product officer Jeetu Patel said last month on the CXOTalk podcast. The company initially gave coding teams months of unlimited tokens to learn the tools; now, managers across his 32,000-person team are to receive token budgets. JPMorgan imposed a monthly $2,000 cap on some engineers’ use of Anthropic’s Claude Code in September, Business Insider previously reported. A bank spokesperson said costs were being tied to productivity, quality, risk reduction and business impact.

The shift follows a period when some companies highlighted high usage. Visa said monthly token use rose from 1 trillion in February to nearly 2 trillion in March. Nvidia CEO Jensen Huang said in March that the company was trying to spend as much as $2 billion a year on tokens for engineers. Companies also used dashboards to track activity: KPMG US has since ended monitoring through its advisory-business dashboard, saying staff now use AI in their everyday work.

Spending and security are driving tighter oversight. Menlo Ventures estimated enterprise AI deployment spending reached $177 billion by September, compared with $33 billion in all of 2025. Yet only 13% of about 2,000 technology executives surveyed by Gartner from May through July said their organizations had gained significant value from AI. A 2025 KPMG survey of more than 1,000 US workers found 44% used AI at work without authorization or in inappropriate ways, and 46% had uploaded sensitive company information or intellectual property to public AI platforms.

Rippling introduced monthly caps in August after projecting token spending could reach 40% of its R&D head count budget. The company said caps and more efficient model routing lowered that projection to 10%–15%, while token use stayed near its previous peak. “The point isn't to cut AI costs,” chief product officer Matt MacInnis told Business Insider. “We're actually fine with cost growth, provided it's productive.”