Companies that spent the past few years encouraging employees to adopt AI are moving to tighter, more selective controls. On November 11, 2026, IT Home reported, citing Business Insider, that Goldman Sachs, Cisco and JPMorgan are among firms setting usage budgets, tracking how often tools are called, and restricting access for particular employees or products. The shift comes as they increase AI spending while trying to reduce related security risk.
Lari Hämäläinen, a senior partner in McKinsey & Company’s technology, media and telecommunications practice, said more companies see employee AI use “rising sharply” and therefore need clear rules and limits.
Not every role gets the most advanced model
Goldman Sachs chairman and CEO David Solomon said this month that the bank is rapidly expanding AI tools so its 45,000 employees can use more products, but that not every employee needs daily access to the most advanced models.
In a Bloomberg Television interview, he said some work does require frontier models, “especially cybersecurity,” while “many jobs do not need the most advanced models.” Ordinary AI tools, he said, are enough to help those employees work more efficiently.
From open use to token budgets
Other companies that first used relatively loose policies to get staff using AI are now limiting the related spending.
Cisco chief product officer Jeetu Patel, speaking last month on the CXOTalk podcast, compared the company’s early approach to learning to ride a bicycle. Cisco allowed development teams to use tokens for AI coding tools without a limit for several months, so employees could learn the technology before efficiency was the priority.
As use grew, Patel said he plans to set token budgets for managers at each level in a team of 32,000 employees.
Business Insider previously reported that in September 2026 JPMorgan limited some engineers’ use of Anthropic’s Claude Code to $2,000 a month. IT Home noted that this was about 13,376 yuan at the exchange rate it cited.
A bank spokesperson said the company is linking AI usage cost to productivity, work quality, risk reduction and actual business impact.
Only months after firms highlighted heavy use
The tighter management comes less than a few months after some companies publicly emphasized how many tokens employees consumed.
Visa said monthly token use rose from 1 trillion in February 2026 to nearly 2 trillion in March 2026, and that it rewarded the employees who used AI the most.
In March 2026, Nvidia CEO Jensen Huang said the company planned to spend up to $2 billion a year on tokens for engineers — about 13.376 billion yuan in IT Home’s conversion. He joked that he would feel “very worried” if an engineer paid $500,000 a year, about 3.344 million yuan at that conversion, used less than $250,000 of tokens, about 1.672 million yuan.
Some companies also launched dashboards so management could see AI use more directly, and sometimes to encourage more of it.
Disney’s AI-use dashboard, introduced in April 2026, showed that the employees using AI most often averaged hundreds of requests a day, with token consumption in the tens of millions.
In May 2026, KPMG US said it hoped a dashboard for its consulting business would encourage about 10,000 employees to use AI more often and more deeply. As staff brought AI into daily work, KPMG stopped monitoring usage through that dashboard. A spokesperson said the tool had finished its original task of promoting adoption.
Chasing value per token
With AI spending still rising, companies are shifting attention from wider use toward return on the investment.
Research from Menlo Ventures, as reported by IT Home, found that enterprise spending on AI deployment reached $177 billion by September 2026, about 1.18 trillion yuan at the rate IT Home cited. For all of 2025, the figure was $33 billion, about 220.696 billion yuan in the same conversion.
Higher spending has not clearly produced matching returns. A Gartner survey of about 2,000 technology executives, conducted from May to July 2026, found that only 13% said their organizations had gained significant value from AI.
Alex Sobol, co-founder of Millennium Alliance, an invitation-only forum for corporate executives, said enterprise AI bills are swelling partly because employees lack an incentive to control use. Without limits, he said, staff “will find every way to use AI.”
Used improperly, broad AI access can also create security problems. A 2025 KPMG survey of more than 1,000 U.S. employees found that 44% had used AI at work without authorization or inappropriately, and that another 46% had uploaded sensitive company information or intellectual property to a public AI platform.
HR-software startup Rippling took a similar step. In August 2026 it set monthly caps on AI tools, after projecting that token spending could reach 40% of the R&D team’s overall labor-cost budget. Rippling said that usage limits and changes to how model calls are routed lowered the projected token-spending share to 10%–15%, while token volume stayed roughly at the earlier peak.
Chief product officer Matt MacInnis said the aim is not to cut AI cost: “As long as this spending can produce real production value, we can fully accept costs continuing to grow.”
Original report: IT Home, November 11, 2026.
