Who did their homework? Pressure? New WWII movie? True story? If not, stop, go watch then come back. Trust me it will be worth the wait. :)
I’ve spent 35 years analyzing how complex systems fail.
I’ve built and bypassed countless firewalls, hunted billions of pieces of malware (and their authors), and founded and run numerous cybersecurity companies (Foundstone, Cylance, McAfee/Intel CTO Office, Qwiet AI) to predict and prevent digital catastrophes before they happen. But after starting and leading multiple companies and looking back at the wreckage of countless failed strategies, I’ve come to one hard realization:
The most complex, volatile, and critical system of all isn’t made of source code or silicon. It’s the human system.
Remember the movie Pressure that I suggested you watch?
Look at the story of British Meteorologist, James Martin Stagg and General Dwight D. Eisenhower’s decision to launch the Normandy invasion. He had the 1944 equivalent of massive data analytics. But his real problem wasn’t the weather data. It was the humans giving it to him. Ike was presented with diametrically opposing forecasts from two brilliant meteorologists (UK Group Captain James Stagg and the American meteorologist team led by Irving P. Krick) and needed to make the D-Day invasion decision alone. The fate of the war, the free world and over 300,000 troops depended on Ike and his gut decision to trust (or not) an unfamiliar meteorologist that Churchill called his best. He needed to make a near split second decision: to trust an unfamiliar voice or his trusted team to predict the forecast and inevitably the future.
One was overly optimistic (Krick), anchored to historical analog data models (the past), brimming with confidence from guiding Ike in prior engagements very successfully and eager to please. The other (Stagg) was cautious, highly analytical and reading the real-time chaos of the atmosphere (the present), and trusting in his “gut” as prediction which was (as is even today) pure guesswork based on statistical correlation not necessarily causation.
But the problem Ike was solving for wasn’t as simple as which weather forecast to model. He had to model the people. He had to weigh their cognitive biases, their native traits, communication styles, and the sheer psychological pressure of the moment. The ultimate decision, perhaps the most consequential of the 20th century, was driven by human context, not just barometric pressure and wind speeds and moisture.
Don’t believe the story? Read his story and watch the new movie Pressure.
The State of Decision Intelligence: Perfect Models, Flawed Realities
Gartner just this year came out with a Magic Quadrant (MQ) for Decision Intelligence (DI) platforms. By all measure, today’s enterprise Decision Intelligence (DI) platforms are incredible feats of engineering. They ingest massive datasets and run predictive analytics to map out the business environment with staggering precision.
But they have a glaring, fatal blind spot.
They treat the organization as a frictionless emotionless machine. They assume that if an algorithm outputs the mathematically “optimal” choice, the humans in the boardroom will rationally accept it, flawlessly execute it and “own” it to the bitter end. They meticulously model the business environment but they almost completely fail to model the decision-makers themselves.
As a result, these DI platforms while highly complex systems regress to the mean because they consider neither the originator (the human) nor the complexity of the required unique and novel outcome.
Like it or not, Humans will be an essential part of all companies of the future for countless reasons but the least of which are “real world” experience (AI does not have this “yet”), wisdom, accountability and novelty (to name a few). Misunderstandings, cognitive biases, impulses, instincts and conflicting work styles are all the zero-day threats of human interaction. And our current DI tools are completely blind to them.
The Exposure Trap: AI as an Accelerant and an Auditor
There is a dangerous assumption in modern enterprises that algorithmic processing naturally filters out human error. It does not.
AI is an accelerant, but it is also a ruthless auditor. It assumes the inputs it receives are objective (just as they were written down and digitized in black and white for the AI pipeline to ingest). When AI processes data generated in toxic, siloed, or fear-based environments, it doesn’t correct the human flaw, it institutionalizes it. And ultimately, AI doesn’t just scale weak decisions faster, it actively exposes them. When a mathematically perfect AI model fails spectacularly in execution, it strips away the camouflage of “market shifts” or “bad luck” and leaves only one core culprit: a broken human decision-making architecture.
Consider the collapse of Nokia’s smartphone dominance. Nokia didn’t lack market data. They lacked psychological safety. Middle managers, terrified of an intimidating executive team, systematically filtered out pessimistic data and fed overly rosy scenarios upward. If you apply advanced AI to a culture of fear, the system simply optimizes a strategy based on a lie. When we fail to model the human dynamics behind the data, the data itself becomes compromised.
The Friction of Reality: “What the Numbers Say” vs. Human Cognition
Data doesn’t make decisions. People do. And they are the only ones that can “own” it.
A predictive model represents a sterile, frictionless reality. But in the real world, decisions have to survive the intense friction of human psychology (the egos, the ingrained cognitive biases, and the collective blind spots of a leadership team). The most mathematically sound strategy is entirely useless if the minds receiving it are biologically hardwired to reject it.
Blockbuster didn’t pass on buying Netflix for $50 million because they lacked data availability. They passed because of cognitive failure. Their leadership had access to the exact same market trends as everyone else, but their decision-making was suffocated by Status Quo Bias and the Sunk Cost Fallacy. They were so anchored to their physical retail model that they rationalized away the threat. A predictive model can highlight the smartest financial move, but it cannot override the cognitive biases of the humans at the helm.
The Missing Layer: Finding the Faulty Latch
The ultimate bottleneck in modern enterprise is no longer computational power. It is human alignment. For too long, we’ve treated human behavior as an unpredictable externality - a.k.a. a “culture” problem. But strategy execution requires alignment, and alignment is behavioral. If we want true Decision Intelligence, we must treat human alignment as an architecture just as critical as our data architecture.
We operate on a single, foundational principle: you cannot separate the decision from the people making it. We must bridge the gap between raw analytics and human reality. By analyzing behavioral insights, working styles, and cognitive biases, you doesn’t just model the optimal business decision, it models the team executing it. And predicts where communications will break down, identifies collective blind spots before they derail a project, and ensures the psychological safety required for honest data.
You cannot separate the decision from the people making it. We must model both. Together.
Just like the meteorologists that made that critical launch decision some 82 years ago to predict the future, we have to stop making decisions by looking in the rear-view mirror while driving. Looking back without looking forward means we end up chasing bad decisions after the fact instead of preventing them by understanding the complex human decision making system.
The future isn’t about replacing human judgment with AI. It’s about using AI to deeply understand and elevate human cognition and collaboration, and finally model the entire room including the humans. After all, they will always have the final call… alone and together.