Companies Push Employees to Use AI — Just Not Too Much

(Bloomberg) — Big companies are encouraging workers to use artificial intelligence. Or they guard access. This stance is changing as corporate employers wrestle between justifying their investments in AI and limiting use as computing costs rise.

Walmart Inc. has restricted employees from using its in-house artificial intelligence agent that helps with tasks at work, according to people familiar with the decision. And Uber Technologies Inc. limits each employee’s monthly spending on certain AI coding tools to $1,500 per tool. The ride-hailing company had already exhausted its annual budget for Claude Code, a popular tool from Anthropic PBC.

Across industries, companies are consuming so many tokens (the basic unit of measurement for AI computing) that some have had to reverse their previous anything-goes stance. This is in sharp contrast to previous messages that either encouraged or outright pressured corporate employees to boldly try various in-house and third-party AI tools. Firms including Accenture Plc and Coinbase Global Inc. have told employees that failing to embrace technology could jeopardize their career advancement and even their jobs. At Starbucks Corp., a quarter of bonuses paid to tech workers will be pegged to departmentwide goals that include the adoption of artificial intelligence.

Sundar Pichai, CEO of Google parent company Alphabet Inc., recently stated that monthly usage of the company’s artificial intelligence products increased sevenfold last year to 3.2 quadrillion tokens. “We’ve heard from many companies that they’ve already exhausted their annual token budgets, and we’re only in May,” he said at the company’s Google I/O conference on May 19. Google’s success has made its computing resources so valuable that its own AI researchers sometimes have to queue.

“Companies are pushing employees to adopt AI in the right way, and rising token costs are a feature, not a bug,” said Matt Kropp, chief technology officer of Boston Consulting Group’s BCG X division, which helps clients implement AI. “However, few companies yet know how to budget for AI, and employees are still learning how to use these tools effectively, so there is definitely waste.”

Amazon Inc. has attempted to minimize so-called “tokenmaxxing” by shutting down a leaderboard that tracks AI token usage because it encourages some staff to perform tasks that don’t solve problems but help employees climb the ranks, according to Business Insider. “Please don’t use AI just for the sake of using AI,” Amazon senior vice president Dave Treadwell reportedly told staff. Salesforce Inc., owner of workplace productivity platform Slack, said in a webinar to customers last week that it is tracking how tokenization creates real business value.

While AI can quickly increase the productivity of individuals and small teams (at McKinsey & Co., staff save 30% of the time they previously spent collecting and analyzing data), it’s a much bigger step to deliver company-wide savings. Most large companies are struggling to justify their ballooning investments with real productivity gains, according to a new global survey of large companies from Bain & Co. The result is a confusing minefield that employees can have a hard time navigating.

“Companies are directing their employees to see who can use AI the most, without a clear plan as to why,” said Siobhan Savage, co-founder and CEO of Reejig, which makes software to manage workflows. “Just ‘Increase usage.’ Some of these companies are now stepping back and creating some whiplash. “If I were an employee, I would be questioning the leadership.”

What being AI ready means across many companies can vary and is often unclear. Most of the time, this just means shuffling coins.

“Everyone is still trying to figure this out,” said Helen Russell, chief people officer at software developer HubSpot Inc. “The playbook that we have adopted, and which I would tell everyone to use, is more like a carrot rather than a stick approach.”

But as spending increases and fear of missing out on AI’s promise spreads, many leaders have thrown away the carrots. “AI is coming for your jobs,” Micha Kaufmann, CEO of freelance marketplace Fiverr, said in a memo to employees, warning that those who fail to embrace AI risk becoming obsolete. Amazon.com CEO Andy Jassy said employees who “become AI-savvy” will “be in a good position,” implying that those who don’t will not.

Nearly 94% of senior executives plan to continue investing in AI in 2026, even if it doesn’t pay off, according to a BCG survey published in January. But companies may need to monitor usage more closely, especially as budgets shift toward agency AI designed to perform business tasks. Agentic AI is “unprecedentedly expensive” and consumes thousands of times more tokens than core tasks as agents continue to absorb computing power in the background, according to new research from a team of academics working with researchers from Google and Microsoft Corp.

This may lead more companies to take a Goldilocks approach to adoption – not too little, not too much. But BCG’s Kropp said pushing back on the use of AI at a time when employees are becoming accustomed to its use could be “defeating the purpose” and limit the benefits companies will ultimately reap from AI.

Phil Kirschner, a workplace consultant who teaches corporate employees how to use artificial intelligence in their jobs, said any change in workplace policies has the potential to be a major blow to employees. But as concerns grow about AI’s impact on jobs, changes to the rules around AI adoption “pose an existential threat.”

–With help from Natalie Lung.

More stories like this available Bloomberg.com