Companies are rethinking their approach to AI tool access for employees, shifting from broad adoption to more targeted deployment. Firms like Goldman Sachs, Cisco, and JPMorgan are implementing strategies such as setting budgets, tracking usage, and restricting access to certain tools or employees. This move comes as businesses pour significant sums into AI while seeking to mitigate security risks and optimize spending.
Goldman Sachs CEO David Solomon noted that while the bank is expanding AI tools for its 45,000 employees, not all roles require the most advanced models. Cisco's chief product officer, Jeetu Patel, compared the company's early AI adoption to learning to ride a bicycle, with coding teams initially given unlimited tokens before focusing on efficiency. JPMorgan imposed a $2,000 monthly cap on some engineers' use of Anthropic's Claude Code, tying costs to productivity and risk reduction.
Rising AI spending is prompting companies to focus on maximizing their investment. A study from Menlo Ventures found that enterprise AI deployment spending reached $177 billion by September, up from $33 billion in all of 2025. However, only 13% of tech executives polled by Gartner reported significant value from AI. Security risks also play a role, with a 2025 KPMG survey finding that 44% of workers used AI without authorization or inappropriately.
Rippling, a workforce-management software startup, set monthly caps on AI tools after projecting token spending could reach 40% of its R&D headcount budget. The company reported that limits and more efficient model routing lowered that projection to between 10% and 15% while maintaining near-peak token use. Matt MacInnis, Rippling's chief product officer, emphasized that the goal isn't to cut AI costs but to ensure productive spending.