Accenture has introduced a monthly cap on employee access to Claude — the AI assistant developed by Anthropic — placing the consulting firm among a growing number of technology-sector companies managing the financial pressures associated with operating AI tools at scale.

An internal document headed "Claude Enterprise Users: Important Updates to Licensing and Usage Costs," reviewed by a television news outlet, outlined that Accenture would cap standard access to Claude Enterprise at $29 (around ₹2,700) per worker each month, with the change taking effect on October 1. Prior to the policy change, a number of staff members had reportedly been drawing on AI token allocations worth upwards of $1,000 a month — around ₹94,800 — to handle workflow automation and software development tasks.

Employees on standard plans now operate under fixed monthly usage caps, whereas a subset of enterprise customers retain access to higher-tier arrangements carrying weekly token allowances of up to $250. Those on a standard account face a monthly spending ceiling of $29, the same threshold set for enterprise-tier access.

Any staff member whose usage exceeds the set threshold is required to file a formal Work Breakdown Structure that ties the additional consumption to a named client or an allocated project budget. Under this framework, access rights for generative AI tools are reviewed on a rolling basis, with unused or underperforming allocations redirected toward teams where those tools demonstrate the strongest documented returns.

Responding to questions about the policy change, an Accenture representative said the company's current focus centres on matching tools to specific tasks, with continued investment directed toward AI capabilities, workforce skills, and enablement programmes aimed at delivering greater value for both clients and the organisation itself.

Separately, Uber moved earlier this year to cap each developer's AI coding software usage at $1,500 per month after the platform's full annual budget for such tools ran out within four months. Limits on how freely AI coding assistants such as Claude Code can operate within corporate networks have also been put in place at firms including Microsoft and Meta, with the stated aim of keeping automated token usage in check.

Alongside the usage restrictions, Accenture has expanded its pool of AI and data specialists to nearly 110,000 people, a headcount increase that runs in parallel with the firm drawing on AI-driven efficiencies throughout its managed services operations.

Corporate use of large language models is moving past open-ended deployment into a phase where return on investment and spending discipline are driving decisions about how far the technology is deployed.