It’s a curious thing, the allure of the “expert.” When faced with a thorny problem, particularly in an area where our own footing feels less than solid, the temptation to call in someone with a string of credentials and a compelling pitch can be immense. We crave solutions, and an expert, by definition, is supposed to have them. Yet, over a good few decades in the technology trenches, and coaching businesses through the often-turbulent waters of growth and change, I’ve developed a healthy, if not an absolutely essential, skepticism towards the generic pronouncements of the self-proclaimed expert, a skepticism that feels, if anything, more rather than less warranted in an age when artificial intelligence can generate the confident language of deep expertise entirely on demand.

I recall a particular Business Unit, let’s call them "LogiStream Solutions," a solid team providing specialized software for a rather niche segment of the logistics industry. They were profitable, had a loyal customer base, but their organic growth had hit a plateau. The BU manager, a bright and driven individual named David, was feeling the pressure. Into this scenario stepped a consultant, a "Strategic Synergist" or some such impressive title, armed with glossy brochures showcasing successes in wildly different sectors: a consumer app here, a retail chain there. The promise was grand: a revolutionary go-to-market strategy, leveraging "next-generation paradigms" that would, apparently, unlock explosive growth. The fee, naturally, was commensurate with the grandeur of the promise.

David was intrigued. His team was capable, but this consultant spoke a language of certainty that was hard to ignore. When I sat down with David, my approach wasn't to dismiss the consultant out of hand. After all, I’ve learned a great deal from genuine specialists throughout my career, the kind of deep, narrow expertise that comes from years of focused effort, often nurtured within the rigorous environment of our universities, for which I hold the utmost respect. The kind of expert who can dissect a specific technical challenge in compiler design or a particularly knotty piece of international tax law is invaluable. This, however, felt different.

My questions to David were simple, rooted in what a physicist might call first principles:
"What, precisely, is the fundamental problem this strategy addresses for LogiStream's customers, in their specific slice of the logistics world?"
"What data do we have (from our own sales records, our customer support logs, our win/loss analyses) that supports the assumptions this strategy is built on?"
"Instead of a big bang, what small, low-cost experiments could your own team devise and run in the next quarter to test one or two of these 'paradigms'?"
"How do these broad-stroke retail or consumer tactics translate to the long sales cycles, the deeply entrenched relationships, and the regulatory complexities that define your market?"

The "Strategic Synergist" had offered a pre-packaged answer. The problem was, it wasn't necessarily an answer to LogiStream's specific question. It reminded me of Richard Feynman’s profound intolerance for what he termed "cargo cult science": activities that mimic the form of scientific inquiry without the underlying honesty and rigorous self-criticism. Feynman, a physicist whose clarity of thought I’ve long admired, understood that the first principle is that you must not fool yourself, and you are the easiest person to fool. Experts, especially those selling complex, hard-to-measure services, can sometimes inadvertently (or advertently) help you do just that. It is a trap that, I would suggest, the current generation of AI tools is not immune to: the fluency of a well-trained model can be remarkably easy to mistake for the earned understanding it merely resembles.

Earned Insight

To their credit, David and his team decided to press pause on the expensive engagement. Instead, they reallocated a small portion of their budget and, more importantly, their team's intellectual horsepower. They dug into their own data with renewed focus. They initiated a series of candid conversations with a dozen of their longest-standing clients and a few who had chosen a competitor. They didn't unearth a silver bullet. What they found were a couple of specific, unarticulated frustrations their current offering didn't fully address for a growing sub-segment of their clientele, perhaps those operating across different regulatory zones in Europe and North America.

They then launched a modest, targeted initiative, not a "revolutionary paradigm," but an enhancement to an existing module and a series of very practical webinars addressing these newly understood pain points. The result wasn't the "explosive growth" promised by the consultant. It was, however, a tangible uptick in engagement from a key customer segment, a couple of significant upsells, and, crucially, a much deeper, internally-held understanding of their market dynamics. They had, in essence, "earned" their insight and the subsequent modest, but very real, growth. It wasn't outsourced; it was cultivated.

The paradox, as I see it, is this: the more acute the problem, the greater the desire for an expert to provide a swift solution. Yet, it's precisely in these moments that uncritical reliance on external pronouncements can be most perilous. True expertise, particularly in the nuanced world of business strategy or market engagement, is so often context-dependent. It's "earned" through the iterative process of trying, failing, learning, and adapting within that specific environment. It's a secret your own team uncovers through diligent, honest inquiry, backed by data, not a generic formula imported from elsewhere.

The Same Paradox With AI

I find myself returning to this same principle when colleagues ask what I make of the current wave of AI tools now available to business leaders. There is, I think, a version of the same paradox at work. Large language models are trained on staggering volumes of text: the accumulated output of countless industries, markets, and management theories, and their outputs can feel, at first encounter, almost indistinguishable from genuine expertise. The confident synthesis, the well-structured recommendation, the apparently comprehensive analysis. Yet that very comprehensiveness is, in a sense, the tell. A model trained on everything cannot have been forged in your specific market, your specific customer relationships, or the quiet institutional knowledge your team has accumulated over years of operating in that particular corner of the world. The questions I once put to David are no less necessary when the advice originates from a machine: what specific problem does this address for your customers, in your particular slice of the market? What does your own operational data say? Fluency of output is not the same thing as depth of understanding, and mistaking one for the other is, I would suggest, a rather expensive lesson to learn.

Fluency of output is not the same thing as depth of understanding.

This isn't to say that all external advice is without merit. Far from it. A fresh pair of eyes, particularly from someone with deep, verifiable experience in a directly comparable situation, can be incredibly valuable. But the responsibility for critical thinking, for testing assumptions, and for owning the outcome, must always remain internal. Outsourcing that core function is a path fraught with risk.

In the complex dance of running and growing a business, particularly in the ever-shifting landscape of technology, a healthy dose of skepticism isn't cynicism; it's a prerequisite for sound judgment. It’s the filter that helps distinguish genuine insight from sophisticated-sounding fluff. And more often than not, the most profound and actionable "expert" insights are already present, latent within your own team, waiting to be uncovered through rigorous, first-principles thinking and a willingness to engage honestly with the data your own operations generate. That, to my mind, is a far more reliable path to sustainable, long-term value.