Artificial intelligence has become the centerpiece of nearly every business strategy, policy discussion and product roadmap. Seemingly overnight, every service is “AI-enabled,” every piece of software “AI-powered,” and every plan includes an “AI strategy.”
Yet for all the excitement, we’ve been here before. Each generation of technology comes with inflated expectations and costly disillusionment. Decades ago, companies mistook digitization for automation. Later, they confused reporting for analytics. Today, they’re rebranding old automation methods as AI. The result is the same: overpromising, overspending and underdelivering.
This isn’t a technology problem: It’s a discipline problem — one we’ve seen before.
Mislabeling automation as AI
In the 2010s, true AI innovation was already underway, though largely invisible. Companies like Amazon and Netflix quietly used advanced machine learning to make astonishingly accurate predictions about customer behavior. Amazon’s systems could anticipate what products a customer might buy next and pre-position them in nearby fulfillment centers. Netflix’s recommendation engine used predictive models to personalize viewing experiences. These weren’t flashy consumer apps, but they created enormous value through smarter operations and data-driven foresight.
Then came late 2022, when ChatGPT brought AI into the mainstream. For the first time, consumers could see and interact with an AI that felt intelligent. The public fascination quickly spread to the corporate world. Boards began mandating “AI strategies.” Executives were tasked with producing fast results. And, in the scramble to show progress, many organizations simply relabeled existing automation as “AI.”
In practice, most of these projects combine legacy automation tools with a large language model (LLM) bolted on for window dressing. They’re built on outdated processes and brittle data, just wrapped in a new interface. Companies are grafting AI onto legacy processes instead of redesigning how those processes should function in an AI-first world.
Automation brings efficiency and consistency, but it’s not intelligence. True AI systems learn, adapt and reason through ambiguity without being explicitly reprogrammed.
That’s the difference between traditional automation and what I call “intelligent automation“: systems capable of handling novelty. Older robotic process automation tools, for example, would crash if a button moved or a data field changed. Intelligent systems can infer the right response and keep working.
This distinction matters. When companies mislabel a rules engine as AI, they inflate expectations and erode trust. Beyond failed projects, the real risk for leaders is loss of credibility before true transformation begins.
A familiar pattern
This cycle of mislabeling is nothing new. Each technological wave has followed the same arc: new capability, inflated promises and disappointing returns.
In the early 2000s, organizations replaced paper forms with web forms and called it automation. The process still depended on people typing in fields; it was digitization, not automation. A decade later, companies adopted visualization tools and called the output “analytics.” One colleague of mine with an advanced degree in business analytics quit her “Data Scientist” role after realizing her job was just building dashboards.
Now we’ve arrived at the AI phase of this same pattern. Each time, the label outpaces the substance, and the result is investment without transformation.
The mirage: When foundations fail
Worse than mislabeling, the current hype distracts us from fundamentals. A CFO I know recently shared that her biggest frustration wasn’t AI or automation at all. It was that core IT systems still fail to deliver on decades-old promises. She traced the problem back to stubbornly bad data, fragmented legacy systems and broken processes. A 2024 Forrester study found that 68% of organizations face data quality and integration challenges, limiting AI success. Gartner predicts that 30% of generative AI projects will be abandoned after proof of concept by the end of 2025 due to poor data quality, inadequate risk controls, escalating costs or unclear business value.
Technology amplifies strengths and exposes weaknesses. When leadership treats AI as a race, teams end up automating bad processes instead of reimagining them.
Five disciplines for real AI value
Breaking the cycle requires discipline. To turn AI from hype to business value, organizations must do five things differently:
1. Define precisely. Create shared, organization-wide distinctions between automation, analytics and types of AI (e.g., machine learning, LLMs, agents). Precision in language drives precision in investment.
2. Anchor to business outcomes. Every AI project must answer two questions: “What decision does this improve?” and “What measurable result will it deliver?” If it can’t, it’s not ready.
3. Fix the foundations. High-quality data, strong governance and integrated systems are essential enablers. You can’t build an AI castle on a foundation of sand.
4. Reshape the culture. AI success won’t come from top-down mandates but from empowered teams. Employees must see AI as an indispensable tool to enhance the firm’s competitiveness, as well as their individual value. Organizations that immediately convert efficiency gains into headcount reductions will stymie progress, because employees will not innovate themselves out of a job.
5. Invest in capability. The future belongs to companies that grow human capital to wield new digital capabilities. Build digital mindset, innovation skills and change management so employees can apply AI continuously and creatively.
Get it right this time
AI isn’t magic: It’s math, data and discipline. The opportunity lies not in chasing the next model release, but in rethinking how decisions are made and work gets done.
We’ve seen this story before with digitization, automation and analytics. Each promised transformation fell short when organizations mistook buzzwords for strategy. Let’s not make the same mistake again.
If we pair today’s powerful tools with clarity, rigor and humility, we can finally turn hype into real progress and avoid repeating the costly mistakes of the past.
Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.