In 1999, you could add ".com" to your company name, write "internet" twelve times in your pitch deck, and walk out of a boardroom with a check. The business didn't need to make sense. It just needed to sound like the future.
We're doing it again. Different word. Identical energy.
The word is AI. And right now, every company — in every industry, in every meeting room on earth — is asking some version of the same question: "How do we add AI to this?" Not do we need AI? Not what specific problem would AI actually fix? Just — how do we get some of that in here before someone notices we don't have it.
We've seen this before
The dot-com bubble of the late nineties was a masterclass in collective delusion. Companies that had never made a dollar were valued at billions because they had a website and a vague promise about "disrupting" something. "First-mover advantage" was the justification for burning cash at a speed that would embarrass a casino. And when the music stopped in 2000, most of those companies had nothing behind the pitch — not a product, not a customer, not a plan.
The AI gold rush has the same mechanics. Companies are bolting AI chatbots onto their websites not because customers asked for one, but because a competitor did it and now they feel exposed. Retailers are building "AI-powered recommendations" on top of five-year-old inventory systems that can barely run a report. Businesses are putting "AI-powered" in their marketing and then doing the actual work manually behind the scenes.
Nobody is stopping to ask whether any of this is actually better than what they had before.
The FOMO tax
Here's the honest reason this is happening: businesses aren't adding AI because they've identified a real problem it would solve. They're adding it because they're terrified of being the company that doesn't have it.
Nobody wants to sit in a client meeting and be the person who "doesn't do AI." Nobody wants to look like they missed it. The fear is understandable — the problem is that "we have AI" and "AI is making us better" are two completely different statements, and right now the market is only asking for the first one.
So companies pay the FOMO tax: they spend money on tools they don't fully understand, for problems they haven't clearly defined, measured by metrics nobody agreed on in advance. Then they call it innovation.
A company that adds an AI chatbot to deflect customer service calls — without fixing the reason customers are calling in the first place — hasn't improved anything. They've just added a layer of automated frustration between the customer and the actual problem. The customer is still angry. They just had to explain themselves to a robot first.
Adding AI to a broken process doesn't fix the process. It just makes the broken parts run faster.
The dot-com parallel holds up here too. In 1999, the fear was the same: be online or be irrelevant. Some companies built real things. Most just bought domain names and called it a strategy. The ones who survived were the ones who had an actual reason to be there.
Where AI actually earns its place
AI is genuinely useful. That's not the argument here — the opposite overcorrection is just as wrong.
The places where AI creates real value are specific. It's good at processing high volumes of repetitive work that would otherwise require human time and attention. It's good at finding patterns in data no human would have the patience to analyze manually. It's good at handling the predictable, routine parts of a workflow so your team can focus on the parts that actually require judgment.
What it's not good at: being the answer to a question nobody asked.
If you're thinking about adding AI to your business, start here — not "where can we use AI?" but "what is slow, repetitive, or expensive in our current operation?" If AI is genuinely the right tool for that thing, build it. If the answer is "nothing obvious comes to mind," then you don't have an AI problem. You have a strategy problem, and AI won't fix it.
The companies that will actually benefit from this era are the ones treating AI like any other tool: useful for specific jobs, pointless for others, and always secondary to understanding the actual problem.
