- Technical roles get $1,000/month and non-technical $100, with $500 for technical staff in sales and customer success
- The pilot team spent well under budget in month two with no output impact
- Salesforce expects roughly $300M in Anthropic token spend this year, mainly for coding agents
AI CURATED HOT STORY · 11/18/2026
Salesforce Caps AI Token Budget at $1,000 Monthly for Tech Staff
Salesforce has implemented role-based AI token budgets starting this month, capping technical staff at $1,000 per month and non-technical staff at $100, with technical sales and customer success roles at $500. The move directly addresses cases where some employees were spending thousands of dollars monthly on tokens, marking a shift from unlimited usage to refined cost management.
According to internal memos and COO Miguel Milano's directive, the budget system was piloted with one team, where actual spending in the second month was significantly below budget with no impact on output. Employees are also required to match models to tasks, prioritizing Slackbot and Gemini for drafting, summarizing, and research. Salesforce expects to spend about $300 million on Anthropic tokens this year, mainly for coding agents, highlighting AI compute as a major corporate expense.
The policy reflects the common dilemma between promoting AI adoption and controlling costs. By setting budget caps, Salesforce ensures reasonable access to AI tools while avoiding waste, offering a cost-management model for the industry. As AI usage deepens, role-based quota management may become standard, driving AI adoption from extensive to intensive.
Hourly heat curve
10 fully observed hoursThe line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.
Coverage timeline
Total 2 reports · Latest first- IT之家 · 智能时代T2·72 pts
- Rohan Paul (@rohanpaul_ai)T2·58 pts
- Budgets of $100/month for non-technical sellers and $1,000 for technical sales follow some staff spending thousands monthly last quarter
- The COO directs matching models to tasks, starting with Slackbot and Gemini for drafting, summarizing and research
- Enterprise AI spending is shifting from open-ended usage to role-based quotas