A study from the National Bureau of Economic Research recently found that while three out of four businesses are currently using artificial intelligence to boost productivity, most report no positive impact. A minority of businesses, such as Klarna and Duolingo, famously reversed course by replacing customer service staff with bots, only to rehire those employees due to customer backlash. However, for most senior executives, the second wave of massive AI expenditures ranks among their top company priorities for the coming years.
Over the past decade, U.S. businesses invested a total of $471 billion in artificial intelligence. In 2026 alone, just five companies plan to spend $700 billion more: Amazon, Microsoft, Alphabet, Meta and Oracle. This new annual investment level is likely to be surpassed each year for the foreseeable future, reaching $1.1 trillion annually by 2029. And what has this historic investment in the holy grail of technology yielded for U.S. businesses?
- 95 percent of AI investments have seen no return whatsoever to the bottom line.
- 42 percent of companies had abandoned their AI investments by the end of 2025, leading to the loss of thousands of jobs in many cases.
- Google erased $100 billion in shareholder value within hours based on a single hallucinated chatbot response.
Now that the most expensive learning exercise in history has exposed how poorly businesses understand this new technology they’ve been deploying at scale, how can we ensure the next trillions of dollars are spent more wisely? More importantly, how can business leaders leverage this remarkable set of capabilities to create opportunities, fuel growth and enhance the lives of their employees, shareholders and customers?
The road to better starts by recognizing and avoiding the pitfalls of the first AI wave, learning from its warning signs and working backwards from the outcomes that truly matter instead of the bullish promises that led to some of the biggest AI failures in recent years.
The Automation Trap
The pattern is consistent across nearly every industry. A company identifies AI as a priority. Leadership greenlights a budget. The team’s first instinct is to automate what already exists. Take the current workflow, bolt AI onto it and measure how much faster it runs. And often, these companies declared premature victory, measuring cost savings without understanding the full impact of these automations. This trap isn’t unique to AI, and is, in fact, yet another case of history repeating itself.
Every general-purpose technology has followed the same trajectory. Electricity, personal computers, the internet. Adoption comes fast, but meaningful gains come more slowly, if ever. Economists call it the Productivity J-Curve: Output declines before it rises, because real value only comes from redesigning a process around what the new tech enables, not from plugging the new tool into an old process in hopes of simply making imperfect outputs faster and cheaper.
When companies prioritize speed over delivering products and services smarter or better, the results speak for themselves.
What Getting It Wrong Looks Like
Speed without judgment almost always results in failure. When AI produces seemingly refined work quickly, people tend to create more, faster and cheaper, rather than using the technology to produce better inputs that improve performance. Thoughtful strategies vanish because the machine often makes them seem unnecessary. Assumptions go unchallenged.
Volkswagen invested $16 billion to establish Cariad, a software unit aimed at developing a single AI-driven operating system for all 12 of its car brands. The company hired 6,000 employees without a unified plan, decision-making authority or structure beyond the siloed processes each employee carried from their home brand.
Instead of resolving issues for VW, Cariad incurred $7.5 billion in operating losses, delayed flagship Audi and Porsche launches by years, and cost Volkswagen’s CEO his job. Volkswagen then paid Rivian an additional $5.8 billion for access to software meant to produce what Cariad never could.
The key difference? VW was a car manufacturer attempting to become a tech leader by pouring money and headcount into the problem without changing its operational approach. Rivian, a tech company that just happens to build cars, delivered exactly what VW needed at a much lower cost and in less time. This pattern has repeated countless times on a smaller scale across corporate America. Consistently, the pressure to demonstrate returns on sunk costs leads to far more costly decisions than the AI was meant to prevent.
What Getting It Right Looks Like
The few companies generating real returns from AI began with a fundamentally different question. Rather than asking, “What can we do faster, cheaper and with fewer people?” they started by asking, “What decisions are we making badly?”
That reframe changes everything. The companies that master AI in its second wave will use it to understand more, not just to produce more. They will analyze customers at a depth previously impossible, and test strategies thoroughly before allocating budgets. Furthermore, by distinguishing causation from correlation and AI-generated hallucinations, every dollar spent can be linked to a specific outcome rather than a vanity metric on a dashboard.
This is the difference between using AI to save money and using it to create new value. The companies seeing AI returns of 5x to 12x aren’t just creating more content or expanding audience segments. They’re making fewer, smarter decisions and targeting resources at validated, scalable opportunities.
The Untold Risk: Talent Depletion
There is one more consequence of the automation-first approach that every leader should consider. If AI can handle entry-level analytical, creative and operational tasks, and companies stop hiring junior talent as a result, who will become the senior leaders in five years? One can’t automate their way to good judgment. People develop it by doing the work, occasionally making mistakes and learning to distinguish what truly drives impact from most of the AI-generated noise out there. Companies that cut junior staff to save money in the short term have hurt their future leadership pipeline, showing they aren’t committed to investing in talent and fostering the current wave of entrepreneurial competition.
The Next Question for Every C-Suite Leader
The question every leader should stop asking is, “Are we using AI?” You are. Everyone is. The question that separates the next chapter from the last one is harder and more important: “What can we see now with AI that we couldn’t see before, and what are we doing about it?” In five years, when we tell the tales of this second AI wave, the companies that best answered that question won’t have simply survived the AI revolution. They will prove the business case for generations to come.
