AI Productivity Paradox

The AI Mandate

There is a new, slightly menacing message echoing through the halls of corporate America: use AI, or else. The initial, abstract fear that a robot might one day take your job has been replaced by a much more immediate and tangible one: your boss might fire you for not using the robots enthusiastically enough.

This isn't hypothetical. As The Wall Street Journal reported last week, the mandate is coming from the very top. Accenture's CEO, Julie Sweet, recently announced the company is "exiting" employees who aren't getting the hang of AI. Here's the WSJ:

[Accenture] has trained about 70% of its roughly 779,000 employees in generative artificial-intelligence fundamentals, [Sweet] told investors. But employees for whom "reskilling, based on our experience, is not a viable path" will be shown the door, Sweet said.

At another company, IgniteTech, the management required staff to devote 20% of their week to AI experimentation, and proceeded to cut those with the lowest adoption rate:

[IgniteTech] employees self-assessed their AI usage and, afterward, the company used ChatGPT to rank the results. After a human review, IgniteTech cut the lowest-scoring performers.

"By their own admission, they’re in the basement," [IgniteTech CEO] said. "So now they have to leave."

This is the new corporate reality. Performance reviews at firms like McKinsey and KPMG now explicitly measure your "AI will." Job applicants are being quizzed on how they use AI in their personal lives. The message is clear: get on the train, or get left behind.

You can almost understand the business logic, until you look at the numbers. And this is where the story gets really weird. Because while CEOs are demanding that everyone use AI, the employees who are actually using the technology don't seem to think it's very useful.

A recent Gallup survey of workers who use AI on the job found that a measly 16% "strongly agreed that the AI tools provided by their organization were useful for their work." More damningly, a comprehensive review by researchers at MIT of over 300 corporate AI initiatives found that only 5% were achieving any quantifiable value.

And this brings us to the great AI productivity paradox. Companies are beginning to fire people for not using a technology that, in most cases, isn't actually making them more productive.

So what's going on here? One answer is that we're in the middle of a good old-fashioned management hype cycle. CEOs, terrified of being left behind in the AI race, are issuing top-down mandates, and the messy work of actual implementation is an afterthought. As one former executive who was fired for pushing back on his company's "brute force" AI strategy lamented to WSJ, "If your AI plan doesn't work out the way you expected it to, it's a huge risk for the business."

Another, more cynical take is that this is all just a convenient new narrative. As a recent CNBC report explored, companies may be engaging in "AI washing"—using the shiny new technology as a convenient excuse for regular old-fashioned layoffs.

When UPS cut 38,000 jobs, it spun the cuts as an "end-to-end process redesign effort which will align [its] organizational processes to the network reconfiguration." In reality, most of the cuts were linked to closing buildings and rightsizing staff after the pandemic boom. Even Amazon's CEO, Andy Jassy, had to walk back the impression that his company's recent layoffs were AI-driven, claiming "it's culture."

This is the strange, contradictory state of AI in the enterprise. In the C-suite, it's a revolutionary force that justifies radical, immediate change. On the ground, it's a tool that employees are skeptical of and that has yet to prove its worth. And in the middle, it's a powerful new justification for whatever a company wanted to do anyway. The one thing that's clear is that if your boss has a new AI hammer, everything is starting to look like a nail.

More on AI at Work:

• Is Google Trying to Give Everyone Their Own Bloomberg Terminal? (ARPU)
• Cognizant will make Claude available to 350,000 employees (Anthropic)
• The state of AI in 2025 (McKinsey)

On Our Radar

Our Intelligence Desk connects the dots across functions—from GTM to Operations—and delivers intelligence tailored for specific roles. Learn more about our bespoke streams.

OpenAI's Capital Crunch

• The Headline: OpenAI Scrambles to Reassure Market After CFO Suggests Government 'Backstop' for Financing (The New York Times)
• ARPU's Take: This was a public glimpse into the immense financial pressure OpenAI is under. The company is locked in a capital spending war with tech giants, but unlike them, it doesn't have a cash-printing machine in the basement to pay for it.
• The Operations Implication: This incident exposes OpenAI's core operational vulnerability: it is forced to compete in a capital expenditure race against hyperscalers like Google and Meta, but without their vast, diversified free cash flow. While its rivals can internally fund their multi-billion dollar build-outs, OpenAI must rely on complex, external financing, making its operational model inherently more fragile and subject to market sentiment. The CFO's comment was a public admission of this underlying structural disadvantage.

Anthropic Courts Europe

• The Headline: Anthropic Expands European Presence with New Offices in Paris and Munich to Capture Enterprise Demand (Reuters)
• ARPU's Take: This is a targeted strategy to capitalize on Europe's deep-seated anxiety about American AI dominance. Anthropic is betting that a local presence and a "safety-first" message will be the winning formula to capture the continent's industrial giants.
• The Go-to-Market Implication: The GTM move here is to position Anthropic as the "sovereign-friendly" AI partner for the European enterprise market. By establishing a physical presence in key industrial hubs like Germany and France, Anthropic can directly address deep-seated European concerns around data privacy, regulatory compliance (EU AI Act), and over-reliance on US hyperscalers. This creates a powerful competitive wedge against OpenAI, whose aggressive commercial posture is viewed with more suspicion by European regulators and industrial giants.

P.S. This is our public analysis. Our real work connects each headline to its effects across critical business functions. We deliver this multi-lens analysis as a confidential briefing—a focused entry point to our full intelligence desk. Reply with 'Briefing' to scope one.