Accenture will centrally pay $29 per person each month for Claude Enterprise usage, effective October 1, under an internal announcement seen by 404 Media. The consulting firm’s new cap comes as some employees have reported monthly Claude bills of $500 to $1,000, largely from heavy use of Claude Code. Accenture has promoted AI adoption among its clients, while its memo says all Claude costs are now consumption-based.

Employees on the standard General plan who exceed the $29 allowance must have a project or team cover the additional cost through a Work Breakdown Structure (WBS), Accenture’s billing code for a specific project or team. Premium includes $250 of token usage each week. Most employees are on the General plan, according to the report. Licences are tied to role and business need, reviewed periodically, and can be reassigned; providing a WBS code does not guarantee continued access.

Messages viewed by 404 Media show employees reacting to the policy. One said monthly usage had exceeded $500 for several months and asked about moving to Premium, but the upgrade appeared blocked on the support page. Another, who uses Claude Code to build and test an AI accelerator for several Accenture clients, said monthly usage typically reaches $500 to $1,000 and called the change disappointing. A third employee said their access had been revoked. An Accenture spokesperson told 404 Media the company provides AI tools to employees, seeks to match tools to work, and continues investing in AI capabilities and skills.

The report places Accenture’s move alongside other companies’ limits on employee AI spending. Microsoft introduced division-level token budgets in July and changed its internal default to a cheaper model; CoreAI executive vice president Jay Parikh told staff that “tokenmaxxing is not what we are optimizing for.” Months earlier, Microsoft’s Experiences and Devices group cancelled most Claude Code licences and moved engineers to GitHub Copilot CLI. The article defines tokenmaxxing as employees using as many AI tokens as possible to appear productive on usage dashboards. Meta removed its internal “Claudeonomics” leaderboard, which ranked staff by token consumption, within days. Amazon, Uber, Walmart and AT&T have also capped or throttled employee AI spending this year. The article says Accenture’s policy requires teams to account for what token spending delivers.

Accenture will centrally pay $29 per person each month for Claude Enterprise usage, effective October 1, under an internal announcement seen by 404 Media. The consulting firm’s new cap comes as some employees have reported monthly Claude bills of $500 to $1,000, largely from heavy use of Claude Code. Accenture has promoted AI adoption among its clients, while its memo says all Claude costs are now consumption-based.

Employees on the standard General plan who exceed the $29 allowance must have a project or team cover the additional cost through a Work Breakdown Structure (WBS), Accenture’s billing code for a specific project or team. Premium includes $250 of token usage each week. Most employees are on the General plan, according to the report. Licences are tied to role and business need, reviewed periodically, and can be reassigned; providing a WBS code does not guarantee continued access.

Messages viewed by 404 Media show employees reacting to the policy. One said monthly usage had exceeded $500 for several months and asked about moving to Premium, but the upgrade appeared blocked on the support page. Another, who uses Claude Code to build and test an AI accelerator for several Accenture clients, said monthly usage typically reaches $500 to $1,000 and called the change disappointing. A third employee said their access had been revoked. An Accenture spokesperson told 404 Media the company provides AI tools to employees, seeks to match tools to work, and continues investing in AI capabilities and skills.

The report places Accenture’s move alongside other companies’ limits on employee AI spending. Microsoft introduced division-level token budgets in July and changed its internal default to a cheaper model; CoreAI executive vice president Jay Parikh told staff that “tokenmaxxing is not what we are optimizing for.” Months earlier, Microsoft’s Experiences and Devices group cancelled most Claude Code licences and moved engineers to GitHub Copilot CLI. The article defines tokenmaxxing as employees using as many AI tokens as possible to appear productive on usage dashboards. Meta removed its internal “Claudeonomics” leaderboard, which ranked staff by token consumption, within days. Amazon, Uber, Walmart and AT&T have also capped or throttled employee AI spending this year. The article says Accenture’s policy requires teams to account for what token spending delivers.