Last week, a study dropped that should be getting more attention. Researchers surveyed nearly 6,000 CEOs, CFOs, and executives across the US, UK, Germany, and Australia. The finding: 90% said AI has had no measurable impact on productivity or employment at their companies.
They didn’t say, "a slight impact." They said zero impact.
70% of those surveyed are actively using AI. They're spending the money. They're rolling out the tools. They're just not getting results.
I've been thinking about why, and I think the answer is older than AI. About 40 years older.
The Theory of Constraints Explains the Failure
In the 1980s, Eli Goldratt wrote The Goal, a book about manufacturing that's really about how any system works. His insight was simple: every business has a constraint that determines the throughput of the entire operation. Improve anything that isn't the constraint, and you've improved nothing. You've just moved work around.
That's exactly what most companies are doing with AI.
They're giving every department a chatbot. They're automating random processes. They're running pilots in marketing, in HR, in finance — everywhere except the one place that actually limits the business. And then they're surprised when nothing moves.
Why Employees Resist — And Why They're Not Wrong
There’s another aspect to this that makes implementation hard. A third of employees in one survey admitted to actively sabotaging their company's AI initiatives. They refused to adopt tools, input bad data, and tampered with performance metrics.
Among Millennials and Gen Z, it's 41%.
The knee-jerk reaction is to blame the employees. But when you dig into why, the picture shifts. One-third of workers say AI threatens their sense of value and creativity. They're not lazy. They're protecting their identity.
And honestly? When your company rolls out AI on a process that was working fine, with no clear reason besides "we need to be using AI," can you blame them?
The constraint-first approach fixes this. When you start with the bottleneck, the place where work visibly piles up, where everyone can see the pain, the AI deployment isn't threatening. It's relief. You're not replacing what people do. You're taking the pressure off the spot that's crushing them.
Employees resist AI that feels arbitrary. They welcome AI that solves a problem they live with every day.
What You Actually Get: A Signal System for Leadership
There's a third benefit to this approach that most people miss.
As you free one constraint, the bottleneck shifts. You deploy AI to speed up proposal writing and now you don't have enough leads coming in. You fix the lead flow and now you can't hire fast enough to deliver. The constraint moves, and you follow it.
Each time you run the process, you're deploying AI on a different part of the business. And each deployment generates data: what's flowing, what's backing up, what's improving.
Do that three or four times and look at what you've built: a monitoring layer across your entire operation. Not because you planned a big data project or deployed a BI platform but because you built the strategic context layer of your business almost by accident, just by chasing the next bottleneck.
That's Applied Intelligence. You start by fixing the thing that hurts. You end up with a signal system that tells leadership where to look next.
Find Your Starting Point
We're building the system to help companies do this: find the constraint, deploy AI on it, and repeat as the bottleneck shifts.
If you want to see what it looks like for your business, take the AI Bottleneck Assessment.
It takes about 15 minutes. You'll get a Bottleneck Map with specific AI recommendations sent to your email.