Companies force employees to use AI, but not too much

Applications of AI


Large companies are encouraging their employees to use AI. Or gatekeeping access. This stance is evolving as employers struggle between justifying investments in artificial intelligence and reining in its use as computing costs increase.

Walmart has restricted employees from using internal AI agents to assist with workplace tasks, according to people familiar with the decision. Uber Technologies also limits each employee’s monthly spending on certain AI coding tools to US$1,500 (RM5,946) per tool. The ride-hailing company had already exhausted its annual budget for Anthropic PBC’s popular tool, Claude Code.

Across the industry, companies are consuming large amounts of tokens, the basic unit of measurement for AI computing, and some are having to revert to their previous anything-goes attitude. This is in contrast to previous messages that encouraged or directly pressured corporate employees to boldly experiment with a variety of internal and third-party AI tools. Companies like Accenture and Coinbase Global have told employees that not embracing this technology could jeopardize their career development and even their jobs. At Starbucks, a quarter of the bonuses paid to tech employees will be tied to department-wide goals, including AI implementation.

Sundar Pichai, CEO of Google’s parent company Alphabet Inc., recently said that monthly usage of his company’s AI products has increased sevenfold over the past year to 320 trillion tokens. “We’re hearing that many companies have already exhausted their annual token budgets, and it’s only 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 may have to keep pace.

“Companies are doing the right thing to push AI adoption into their workforces, and the increased token cost is a feature, not a bug,” said Matt Kropp, chief technology officer at Boston Consulting Group’s BCG X division, which helps customers implement AI. “However, few companies still know how to budget for AI, and employees are still learning how to use these tools effectively, so there is certainly waste.”

Amazon Inc is trying to minimize so-called “token maxing” by shutting down leaderboards that track the use of AI tokens, which doesn’t necessarily solve the problem but incentivizes some staff to perform tasks that help them move up the ranks, according to Business Insider. “Don’t use AI just for the sake of using AI,” Dave Treadwell, Amazon’s senior vice president, reportedly told employees. Salesforce Inc., which runs the workplace productivity platform Slack, said in a customer webinar last week that it tracks how the use of its tokens creates real business value.

While AI can rapidly improve productivity for individuals and small teams, at McKinsey & Company, we’re saving 30% of the time our employees spend collecting and analyzing data, but it’s an even bigger leap forward to realize company-wide savings. Most large companies have a hard time justifying increased investments with real efficiency gains, according to a new global survey of large companies conducted by Bain & Company. The result is a muddy minefield that employees can struggle to navigate.

“Companies are relying on their employees to figure out who can best leverage AI without having a clear plan for why,” said Siobhan Savage, co-founder and CEO of Liesig, which makes software for workflow management. “It was just, ‘Please increase your usage.’ Some of those companies are now going backwards and having a little bit of whiplash. If I were an employee, I would question the leadership.”

For many companies, what AI-enabled means varies and is often unclear. In many cases, you end up just raking in the tokens.

“Everyone is still figuring this out,” said Helen Russell, chief human resources officer at software developer HubSpot. “The strategy we’ve adopted, which I would tell everyone to use, is a carrot-and-stick approach.”

But as spending increases and fears of missing out on AI’s potential grow, many leaders are throwing away the carrot and looking for the stick. Mika Kaufman, CEO of freelance marketplace Fiverr, told employees in a memo that “AI is coming to take your job” and warned that those who fail to embrace AI risk becoming obsolete. Amazon.com CEO Andy Jassy said employees who are “AI savvy” are “in a good position” and suggested those who aren’t will continue to be.

According to a BCG survey released in January, about 94% of senior executives plan to continue investing in AI even if it doesn’t produce a return in 2026. But as budgets shift toward agent AI specifically designed to perform business tasks, businesses may need to closely monitor usage. Agent AI is “extremely expensive” and consumes 1,000 times more tokens than basic tasks as agents continue to consume computing power in the background, according to a new study by a team of academics working with researchers from Google and Microsoft Corp.

This may encourage more companies to adopt a Goldilocks implementation approach. Not too little, not too much. But BCG’s Kropp said encouraging the use of AI when employees are just starting to get the hang of it “could be counterproductive” and could limit the benefits companies ultimately derive from AI.

Phil Kirshner, a workplace consultant who teaches corporate employees how to use AI at work, said changes in workplace policies could give employees whiplash. However, changes to rules around AI implementation “come with an existential threat” amid growing concerns about the impact of AI on jobs. – Bloomberg



Source link