Thinking Harder Is Not the Same as Thinking Better
You know the pose. Leaning forward. Fingers pressed against your temples. Brow furrowed just enough to suggest something profound is happening. From the outside, it looks like strategy. From the inside, it is often the same three thoughts running laps around your brain while you quietly hope one of them develops a personality.
The answer rarely appears when you think harder. It appears when you think differently.
That sounds obvious, but most organizations reward effort more than perspective. When something is not working, we add more. More meetings. More data. More experts. More slides explaining the same problem in increasingly impressive visuals. Eventually the arrows start moving, the boxes get rounded corners, and everyone feels better because the problem now has a gradient.
I have done this myself. Years ago, I was working with a manufacturing team that wanted to improve throughput. The conversation immediately went where these conversations usually go: faster machines, better scheduling, more automation, tighter KPIs. We spent a lot of time trying to optimize the line. Then someone asked a much less sophisticated question: why are we producing this batch size in the first place?
That question wrecked half the analysis, which was annoying because it was good analysis. It was also the beginning of the answer.
The Trap of Thinking Harder
Harder thinking feels productive because it creates motion. You can build a workstream around it, assign owners, and schedule a recurring Tuesday meeting that everyone will eventually resent. But harder thinking often just takes us deeper into the assumptions that created the problem in the first place. If the original question is wrong, more effort mostly produces a more detailed wrong answer.
This happens constantly in transformation work. A company asks, “How do we automate this process?” before asking whether the process deserves to survive. A leader asks, “How do we overcome resistance?” before asking what employees believe they are protecting. A technology team asks, “Which AI platform should we choose?” before anyone has agreed on which decisions actually need to improve.
I once bought a piece of furniture that came with instructions containing no words, only tiny drawings of a cheerful cartoon person assembling it with impossible confidence. Halfway through, I realized I had installed a panel backward. My response was not to reconsider the orientation. My response was to push harder. This is apparently also how many companies approach digital transformation.
The problem is not effort. Effort matters. Concentration matters. Expertise matters. But they only become useful after we create enough distance from the problem to see the frame around it. Thinking differently usually starts by changing the question. Instead of asking, “How do we make this process faster?” ask why it exists, who depends on it, and what would happen if it disappeared. Instead of asking which technology to implement, ask which decision needs to become faster, better, safer, or more consistent. Instead of asking how to get people to comply, ask what makes the current behavior rational from their perspective.
That last one is uncomfortable. It is much easier to call people resistant than to admit they may understand the operating reality better than the transformation team does.
How to Think Differently With AI
AI can make this problem better or much worse. Used lazily, it becomes an industrial-scale confirmation machine. Give it your assumptions, your preferred answer, and a prompt that quietly points toward the conclusion you want, and it will often return a polished explanation of why you were right all along. Used properly, AI can create cognitive friction. Not an oracle. Not a digital executive. More like an endlessly patient colleague willing to challenge the framing, simulate other viewpoints, and ask the irritating follow-up questions humans tend to avoid after the third meeting.
A few ways I use it:
Ask it to identify the hidden assumptions behind the problem statement, then rank which ones are most fragile.
Ask it to rewrite the problem from the perspective of an operator, customer, competitor, CFO, regulator, and new employee.
Ask for three explanations that contradict one another, then identify what evidence would distinguish among them.
Ask what would need to be true for the opposite strategy to be correct.
Remove the proposed technology entirely and ask it to solve for the same business outcome another way.
None of these prompts are magic. The value comes from forcing movement in the frame. You are not asking AI to give you the answer. You are using it to expose how many different versions of the problem might exist. This is also where technical discipline matters. AI outputs are shaped by context, wording, model behavior, and the information you provide. A confident answer is not a validated answer. I like to separate divergence from convergence. First, use AI to expand the possibilities, challenge assumptions, and create competing hypotheses. Then bring in evidence, operating data, domain experts, and actual constraints to narrow the field.
AI is very good at generating possibilities. It is less qualified to decide which possibility will survive contact with your factory, customers, employees, budget, regulations, legacy systems, and that one machine installed in 1987 that apparently no one is allowed to turn off. In other words, let AI make the room bigger before you start choosing furniture. And yes, I am still thinking about that backward panel.
Better Questions Create Better Options
The biggest breakthroughs I have seen were rarely brilliant new answers. They were old problems viewed from an angle no one had bothered to try. A maintenance problem became a scheduling problem. A technology adoption problem became a trust problem. A productivity problem became a product-mix problem. A data problem became an ownership problem.
Nothing magical happened. Someone simply stopped accepting the original label.
That is the part that matters most in an AI world. AI will make answers cheaper, faster, and far more abundant. It will draft the memo, summarize the research, generate scenarios, write code, and produce a disturbingly polished chart before you finish your coffee. The scarce skill will be knowing what deserves to be asked, which assumptions should be challenged, and when the first answer is merely the most convenient one.
So yes, concentrate. Dig in. Do the work. But occasionally take your fingers off your temples, step away from the problem, and ask whether you are solving it or simply becoming more sophisticated at describing it.
Thinking harder may help you complete the instructions. Thinking differently is what helps you notice the panel is backward.
