According to The Wall Street Journal, three years into the AI boom, corporate America is “all in” but the reality is less exciting than the hype. Reporters note that adoption is largely about automating existing workflows, like using Microsoft Copilot licenses or summarizing unstructured data from emails and PDFs. While companies like Lockheed Martin, Walmart, and Bank of New York are deploying AI, it’s often behind the scenes for steady, incremental payoffs, not transformation. A recent MIT study suggests 95% of AI pilots fail to deliver meaningful value, yet CEO confidence remains high, with benefits often measured in “cost avoidance” rather than direct revenue. The conversation also highlights a cautious approach to “agentic” AI, with companies like Ford emphasizing a human-in-the-loop model, and acknowledges that AI is a factor in stagnant hiring, as seen in earnings reports where firms brag about low hiring alongside AI investment.
The Unsexy Reality
Here’s the thing: the AI revolution in the office looks a lot like… slightly better software. It’s summarizing meeting notes. It’s combing through support tickets. It’s the digital equivalent of optimizing the supply closet. Steven Rosenbush nailed it with the driverless car analogy—initially thrilling, then just boringly competent. That’s where we are. Companies aren’t building Claude-powered super-strategists; they’re installing chatbots that (hopefully) don’t promise customers crazy deals or spout propaganda. The flashy, autonomous “agent” is mostly a marketing term right now, because frankly, do you want a piece of software making multi-day decisions for your whole business? Most companies, wisely, don’t.
ROI, Jobs, and the Vendor Dilemma
So if the gains are incremental, why are CEOs so bullish? I think it’s a long-term belief play. The ROI is in “cost avoidance”—the idea that you’ll only need five new engineers next year instead of a hundred because of AI coding assistants. That’s a powerful narrative, even if it’s hard to pin down on a spreadsheet. And that narrative directly ties to jobs. Look, when companies post record profits while keeping hiring at zero and name-checking AI, it’s naive to think there’s no connection. Is AI the only factor? Of course not. But it’s in the mix, and it’s a convenient scapegoat for Wall Street.
And who’s winning? The big AI model makers are fighting it out, but enterprises are terrified of lock-in. They got burned by the cloud wars and are now determined to be model-agnostic. No one wants to be stuck in a single ecosystem again. So the competitive advantage won’t come from which chatbot API you use.
The Real Challenge Is People
Where does advantage come from then? People. Getting your employees to actually use the tools effectively. This is the hard part. You can buy 10,000 Copilot seats, but if your culture doesn’t encourage experimentation and trust, they’ll gather digital dust. We’ve all had our “magic” AI moment—planning a vacation, building a silly app. But that playful, consumer-grade freedom rarely exists inside corporate firewalls. Companies appoint “champions” and roll out trainings, but the breakthrough happens when they just give people the tools and a sandbox to play in, without a million restrictions.
The Nvidia Exception and The Future
Is anyone doing it right? The WSJ panel points to Nvidia, and it’s a telling example. AI is in their DNA, from chip design to customer service. It works because it comes from the very top—CEO Jensen Huang is a “true believer,” and that curiosity trickles down. But let’s be real: 99.9% of companies aren’t Nvidia. For them, the path is slower. It’s about leadership buy-in, patient investment, and accepting that the near-term wins will be boring. The boring stuff, like automating a report or fielding a customer query, is where the foundation is being poured. The flashy agentic future? It’s waiting for the technology—and our trust—to catch up.
