Goldman Sachs, Cisco and JPMorgan target workplace AI access

The companies’ approaches include setting budgets, tracking usage and restricting access to particular employees or AI tools.[1] The article links that targeting to rising AI spending and efforts to reduce security risks.[1] McKinsey partner Lari Hämäläinen said that as AI use rises, employers need to set guardrails.[1] The practical change is not simply whether a company adopts AI, but which workers can use particular tools and under what limits.[1] Teams planning deployments can treat access rules, usage oversight and risk controls as part of the rollout rather than as separate questions.[1]

Goldman Sachs is rapidly expanding its range of AI tools for its 45,000 employees.[1] Bloomberg reported that the bank restricted most staff from using advanced AI coding tools after reviewing productivity and risk concerns; its internal guidance, updated in October, limited Claude Code and GitHub to engineers.[2] Fast Company reported that traders were frustrated by the restriction after using the tools to monitor data, build financial models and produce graphics.[3] A Goldman spokesperson said staff can experiment with some AI tools to identify use cases and assess productivity and risk, after which tools go to those best suited to use them.[3]

Cisco plans to introduce token budgets for managers after coding teams had months of unrestricted use.[4] JPMorgan has capped some engineers’ Claude Code spending at $2,000 per month.[4] The examples show different ways to constrain use: budgets can limit consumption, while access restrictions can reserve a tool for selected roles.[4] A 2025 KPMG survey found that 44% of more than 1,000 US workers had used AI at work without authorization or in inappropriate ways, and 46% had uploaded sensitive company information or intellectual property to public AI platforms.[4] Those survey results offer context for employer concerns, not evidence about the conduct of staff at these three companies.[4]