{
  "id": 17395993,
  "title": "Our AI agent experience: we scrapped the multiple $60,000 salary marketing jobs and hired one person at a far higher salary",
  "url": "https://urgent.news/2026/10/31/our-ai-agent-experience-we-scrapped-the-multiple-60-000-salary",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-31T11:30:00.000Z",
  "source": {
    "name": "Fortune",
    "slug": "fortune",
    "url": "https://fortune.com/2026/10/31/dont-ration-company-ai-token-budgets-caps-accenture/"
  },
  "original_language": "en",
  "account": "In July, Accenture advised companies on managing their AI expenses. The company recommended avoiding blanket usage cuts, routing work to the cheapest model that fulfills the task, and focusing on the cost of a finished piece of work before and after incorporating AI. According to internal reports obtained by 404 Media, Accenture began centrally funding $29 per month per employee for Claude, with any additional usage charged to specific projects. Internal messages revealed that employees were spending $500 to $1,000 monthly on AI tools, with some having their access revoked entirely. While Accenture denies limiting employees' AI access, the distinction is significant as fixed-price contracts make up more than 65% of the company's bookings, resulting in token costs impacting its margins. However, from an employee's perspective, the distinction is less clear-cut, and Accenture's records indicate a high demand for daily increases in individual spending limits. Companies are grappling with the question of how many tokens to allocate per employee, which is the wrong question. The right question is the cost of a finished piece of work, whether AI is involved or not. A recent example illustrates the impact of AI on hiring decisions. Instead of hiring multiple entry-level marketing employees at around $60,000 each, one non-technical person was hired and paid significantly more. This individual took on tasks that the group could not accomplish, such as designing collateral and building analytics software to measure customer operations against industry peers. While AI is billed by the token, with a million tokens (about 750,000 words or nine novels) costing between $2 and $8, heavy users can spend up to $89 annually on AI usage. This cost is a fraction of the total bill, which is often underestimated, with the token bill accounting for about a quarter of the expenses in several industries, such as bank fraud detection, pharmaceutical research, chemical discovery, and chemical manufacturing.",
  "summary": "Don't ration your employees' AI use. Overspending is recoverable next quarter, but being left behind isn't.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}