Companies Are Scrambling to Stop Spending So Much on AI

6 hours ago

Accenture is struggling with soaring AI token spend, largely driven by non-technical workers using tokens for trivial tasks like converting PDFs to presentation slides, as revealed in leaked audio.
The tech industry is shifting from uninhibited AI growth to cost management, with companies like GitHub charging per token and Uber capping employee AI tool usage after blowing its entire AI budget in four months.
Contrary to the narrative that engineers drive AI usage, internal data shows non-engineers are major token consumers for non-specialized tasks, undermining efficiency.
During an internal meeting, Accenture's Justice Kwak discussed 'token ops,' highlighting rapid escalation in token spend as AI scales, leading to unpredictable costs and leadership questioning ROI.
Token spending is increasing exponentially as AI adoption grows, creating a widespread problem for enterprises bullish on AI, with Accenture planning to address it through 'token economics'.
Accenture's proposed solution includes launching 'Token IQ' to improve token management, but controls like budgeting are arriving too late to curb current overspending.
The company's push for AI adoption includes requiring senior staff to use AI or risk losing promotions, aligning with its strategy to be an AI-enabled, client-focused reinvention partner.
Other companies like Walmart have also capped AI tool usage due to high demand, and some startups boast about spending on AI over human workers, reflecting broader industry trends.

Hasty Briefsbeta