The Decision Was Made Before the Evidence Was Read

A strategic decision can fail even when an organisation has strong data, capable people and advanced technology. This episode examines what happens when executive preference hardens before evidence is genuinely considered.

It explores how confirmation bias, hierarchy and weak decision architecture can turn analysis into a defence mechanism.

TLDR / At a Glance

• Evidence filtered through executive preference
 • Hidden costs of silenced expertise
 • Decision architecture and explicit assumptions
 • Integrated data, judgement and operational knowledge
 • Strategic Intelligence as an organisational capability
 • AI’s role in scaling insight and bias

Better outcomes depend on leaders creating systems where evidence, expertise and constructive challenge shape decisions before valuable options disappear.

Transcript

Copy Transcript Copied

The Decision Was Already Made

SPEAKER_00 0:00

The decision was made before the evidence was read. How one executive decision changed the way I think about data, judgment, and strategic intelligence. A decision that stayed with me.

A Strategic Call That Lingered

SPEAKER_00 0:15

Years ago, while serving as head of AI for a large global company, I became involved in a strategic decision that has stayed with me because it exposed the considerable distance that can exist between an organization possessing intelligence and actually using it. The company had a meaningful opportunity in front of it, the technical capability required to assess it, and a team whose knowledge spanned data, technology, operations, and commercial reality. The available evidence was not perfect because important strategic decisions rarely arrive with complete certainty, but it was sufficient to support a disciplined discussion about the potential value, the risks involved, and the conditions that would need to be created for the opportunity to succeed. The difficulty was not an absence of data or expertise. The organization had a great deal of both, although much of the information was fragmented across organizational silos, stored within aging technology architecture, and interpreted by functions with different incentives, priorities, and definitions of success. Some of the evidence supported the opportunity, some challenged the prevailing executive position, and some highlighted risks that needed to be understood and managed rather than treated as reasons to stop. None of that complexity was unusual, and none of it made the decision impossible. In fact, working through competing evidence, uncertain outcomes, and legitimate differences of opinion is a normal part of responsible leadership. The

When Evidence Stops Mattering

SPEAKER_00 1:48

real problem was that the decision had effectively been made before the evidence was properly considered. I remember one particular moment in the discussion when a piece of evidence was presented that directly challenged the chief executive's initial position. The evidence was not disproved, nor was an alternative interpretation offered that explained why it should carry less weight. Instead, the conversation shifted towards identifying reasons why it should not alter the preferred direction. At that point it became clear that we were no longer examining the opportunity on its merits. We were examining whether the available information could be made to support a conclusion that had already begun to harden. When

Confirmation Bias Takes The Wheel

SPEAKER_00 2:29

evidence became a defense mechanism. There is a considerable difference between beginning with a hypothesis and testing it rigorously, and beginning with a preferred conclusion and searching for evidence that allows it to survive. The first is a legitimate and necessary part of executive judgment because leaders cannot approach every issue without an initial perspective. The second is confirmation bias, reinforced by organizational authority, and once it enters a decision process, it can be difficult for even strong evidence and experienced people to change the outcome. The chief executive had formed an initial view and as the discussion progressed, became increasingly committed to defending it. Information that supported the position was accepted readily, while data that challenged it was questioned, discounted, or treated as less relevant. The original argument rested on assumptions that the evidence did not fully support, but rather than using that evidence to examine whether the assumptions remained valid, the organization began using analysis to protect the original position from scrutiny. This can create the appearance of a rigorous decision process. Because the organization may still produce reports, models, presentations, and recommendations, yet the existence of analysis does not mean that analysis has been allowed to influence the decision. Evidence can be present in the room without being genuinely considered, particularly when people understand that the acceptable conclusion has already been established by the most powerful person in the discussion. I remember my frustration because the team had done the work responsibly. They had gathered the available evidence, involved people with relevant technical and operational expertise, identified where uncertainty remained, and explained both the strength of the opportunity and the controls that would be required to manage its risks. They were not claiming that the proposal was without difficulty, nor were they trying to replace executive judgment with a spreadsheet or a model. They were attempting to give the leadership team a more complete understanding of the situation so that judgment could be exercised with greater clarity and fewer blind spots. However, questions that should have been explored became challenges to be neutralized, while disagreement that should have strengthened the decision was increasingly treated as resistance to it. The discussion moved away from asking what the strongest available decision might be and towards asking how the existing position could be justified. That was the point at which data stopped being used to improve judgment and started being used to protect judgment from challenge. The

The Human Cost Of Dismissal

SPEAKER_00 5:10

cost went beyond the opportunity. The organization ultimately chose not to pursue the opportunity, and while no one can prove with certainty what the alternative outcome would have been, it gave up potential value that the available evidence, professional expertise, and operational experience suggested was within reach. Any claim about a decision that was not taken must acknowledge the limits of the counterfactual, but the quality of the process can still be assessed, and in this case the process did not make full use of the intelligence available to it. The immediate commercial consequence was significant, but the less visible impact on the team was equally important. People had invested time, judgment, and professional credibility in producing a balanced assessment, only to discover that evidence was welcome when it supported the preferred conclusion, and far less welcome when it challenged it. That experience weakened confidence in the decision process because the team could see that careful analysis, practical experience, and constructive challenge might be ignored when they became inconvenient. Organizations often underestimate how quickly people learn from moments like this. When employees see unwelcome evidence dismissed, they begin to understand the unwritten rules governing what can be said, how directly it can be expressed, and whether challenging a senior position carries more personal risk than organizational benefit. Some people stop speaking openly, while others soften their analysis, remove uncomfortable details, or present only the information they believe senior leaders are prepared to hear. A few continue producing excellent work, but gradually detach from the decision process because they no longer believe their contribution will meaningfully influence the outcome. Over time, this changes the quality of intelligence that reaches senior leadership. Reports become more polished but less revealing. Difficult messages are filtered as they move through management layers, and confidence starts to replace candor. Leaders may then believe that the organization broadly supports their view when in reality the organization has simply learned that disagreement is unlikely to be productive. This is how an organization can become rich in data and poor in intelligence. The decision

Data Rich Yet Intelligence Poor

SPEAKER_00 7:32

exposed a wider system problem. It would have been easy to treat the experience simply as a story about one chief executive making one flawed decision, but that interpretation would have missed the more important lesson. The chief executive's behavior mattered, but the wider organization had also created the conditions in which an individual preference could dominate the intelligence available across the system. Data was fragmented between functions, the technology architecture made information difficult to connect quickly, and organizational silos encouraged teams to interpret the same situation through their own local priorities. Decision rights were not sufficiently explicit, assumptions were not made visible early enough, and there was no agreed process for establishing what evidence would cause the organization to reconsider its direction. The result was a decision environment in which hierarchy could settle a disagreement without resolving the underlying questions. The organization possessed intelligence, but it did not operate intelligently. That distinction became important to me because good data does not automatically produce good decisions, just as experienced leaders do not automatically exercise good judgment. Data can be incomplete, selectively interpreted or presented without sufficient context, while experience can become overconfidence when familiar patterns are assumed to apply to circumstances that have materially changed. Intuition remains valuable because experienced leaders often recognize signals before they can fully articulate why they matter, but intuition is not self-validating and should still be tested against evidence, expertise, and alternative interpretations. The mistake was not that the chief executive had formed a gut view. Leaders frequently need to develop an initial position quickly, particularly when time is limited and the evidence is incomplete. The mistake was allowing that initial view to become increasingly immune from challenge because experienced judgment should provide a starting point for inquiry rather than an end point that the organization is expected to defend. This was therefore not only a failure of individual judgment, it was a failure of decision architecture, organizational culture, and information flow, because the system did not create a sufficiently robust connection between what the organization knew and what its leadership ultimately chose to do.

Connecting People Data And Judgment

SPEAKER_00 9:58

Good data and good people are both required. The lesson I took from the experience was simple in principle but more demanding in practice. Good data and good people create the conditions for better decisions, but only when the organization is prepared to listen to both and has designed a process that allows them to influence the outcome. Data without context can mislead because numbers do not interpret themselves, while expertise without evidence can deteriorate into opinion if it is never tested against reality. Experience can help leaders recognize patterns, but it can also cause them to see a familiar answer where the conditions have changed. Technology can identify relationships at scale, but it cannot determine which trade offs and organization should accept or which strategic outcomes matter most. The strongest decisions emerge when evidence, experience, operational knowledge, technical capability, and professional judgment are brought together within an environment where disagreement is used to improve understanding, rather than treat it as a threat to authority. A better process would not necessarily have removed all uncertainty from the opportunity, nor could it have guaranteed success. What it could have done was make the assumptions visible, separate evidence from interpretation and preference, clarify where the data was strong and where it remained incomplete, and establish what would need to be true for the proposed course of action to succeed. It could also have identified what evidence might reasonably cause the organization to change direction rather than allowing the decision criteria to shift whenever an inconvenient finding emerged. The process would also have involved the right people earlier and more deliberately. The technical team understood what the technology could and could not do. Operational leaders understood how the proposed change would affect workflows, customers and employees, and data specialists understood where the evidence was reliable and where caution was required. Business leaders brought a broader understanding of strategy, capital allocation, and the opportunity cost of delay, while those closer to the work understood practical constraints that might never appear in an executive presentation. No single group held the complete answer, but together they could have produced a more complete understanding. The purpose of involving different perspectives should not be to create decision making by committee or require universal agreement, but to ensure that the person accountable for the choice is exposed to the most relevant evidence, experience, and challenge before the decision becomes difficult to reverse. Good data plus good people does not automatically equal success, but it considerably improves the quality of judgment when both are connected through a disciplined decision process.

From Frustration To Decision Intelligence

SPEAKER_00 12:50

From frustration to decision intelligence. My initial response to the experience was frustration, which was understandable but ultimately insufficient. Frustration may identify that something is wrong, but it does not improve the next decision, strengthen the organizational system, or help another leadership team avoid the same failure. I came to a simple conclusion that has influenced my work ever since. Frustration minus action serves no purpose. The more useful response was to examine why organizations with capable people, substantial data, and significant technological investment still make avoidable mistakes. The answer, in many cases, sits somewhere between information and action. A weak signal may be detected but never reach the person with authority. Evidence may be produced but stripped of context as it moves through management layers, and expertise may exist without being connected to the decision at the point when it could make the greatest difference. A leader may receive analysis but fail to create space for meaningful challenge, while a decision may be made and executed without the organization establishing whether the result confirmed or challenged the original assumptions. This is what sharpened my interest in decision intelligence. I began to focus less narrowly on whether an organization had the right data or the most advanced technology, and more on how its consequential decisions were designed. Who had the authority to decide what evidence was required, which assumptions were explicit, and which remained hidden? Who was invited into the discussion, whose expertise was missing, and what would need to change for the organization to revise its position? How would the decision be translated into action? And what feedback would allow the organization to learn rather than simply move on? These questions often reveal more about future performance than the sophistication of the technology itself. A strong model cannot rescue a decision process that is politically closed. A modern data platform cannot compensate for unclear accountability, and an AI system cannot improve executive judgment when leaders use it only to generate evidence in support of positions they already hold. Technology can widen the field of vision, but it cannot force an organization to look at what it would prefer not to see.

Strategic Intelligence As A System

SPEAKER_00 15:16

How the experience led to strategic intelligence. Decision. Intelligence helped me understand the immediate failure, but over time I realized that decisions could not be improved in isolation from the wider organizational system. The quality of a choice depends on how effectively the organization senses change, interprets what the signals mean, brings evidence and expertise together, executes with sufficient clarity, and then learns from the result. This broader understanding eventually led to the ideas behind the executive's guide to strategic intelligence. The book grew from a conviction that organizations rarely struggle because they know nothing. More often, they struggle because the intelligence they possess is fragmented, delayed, filtered, disputed, or ignored before it can influence meaningful action. Strategic intelligence is therefore not simply about collecting more information or improving the sophistication of analysis. It is the organizational capability to identify meaningful signals, interpret their significance, make timely choices, execute with intent, and learn from the outcomes. It depends on data and technology, but it also depends on human judgment, decision rights, leadership behavior, organizational design, trust, and the willingness to confront evidence that may be uncomfortable. The central challenge is not simply to know more, but to create a system in which the right knowledge reaches the right decision at the right time, and in a form that allows the organization to act before its available options narrow. This also means strategic intelligence cannot remain the preserve of executives, analysts, or specialist teams. The capability needs to be developed throughout the organization because signals frequently emerge far from the senior leadership team. And the people closest to customers, operations, technology, and changing market conditions may see important shifts before those shifts appear in formal performance reporting. People at every level need to know how to recognize meaningful evidence, distinguish fact from interpretation, challenge assumptions constructively, and connect insight to action. Executives remain accountable for the most consequential decisions, but they also shape the quality of the intelligence that reaches them. Through their questions, reactions, incentives, and willingness to reconsider a position, they influence whether people communicate openly or learn to filter what they know. Why

AI Makes Bad Decisions Scalable

SPEAKER_00 17:49

AI raises the stakes. This leadership responsibility is becoming more important as artificial intelligence makes it easier to produce persuasive analysis at speed. AI can generate arguments, forecasts, summaries, and recommendations that appear authoritative, even when they are based on incomplete data, weak assumptions, or a poorly framed question. The technology can strengthen analysis, but it can also make confirmation bias faster, more scalable and more convincing. Used well, AI can support strategic intelligence by identifying patterns that might otherwise be missed, comparing a wider range of scenarios, accelerating analysis, and helping leaders test different interpretations. Used poorly, it can provide sophisticated support for a conclusion that was decided in advance, giving weak judgment the appearance of analytical legitimacy. The executive task is therefore not simply to ask whether an AI-generated answer appears accurate. Leaders need to examine how the problem was framed, which data was included, what was excluded, which assumptions shaped the output and whose judgment is represented in the conclusion. They also need to understand whether the system is presenting evidence, inference, or recommendation, because those are not interchangeable. AI can widen the field of vision, but the organization still needs people who understand the context, can recognize when the output does not make sense, and are prepared to raise that concern even when the answer supports a senior leader's preferred position. The more powerful the technology becomes, the more important it is that leaders create a decision environment in which judgment remains open to evidence and accountability remains clearly human. What

How I Would Redesign The Process

SPEAKER_00 19:37

I would do differently now. Looking back, there are several things I would approach differently, not because I believe the outcome could have been guaranteed, but because the quality of the decision process could have been materially improved. I would make the decision architecture explicit much earlier by clarifying who held authority, which criteria would be used to assess the opportunity, what assumptions required testing, and what evidence could reasonably cause the organization to change direction. I would also bring the conflicting data into one shared view rather than allowing separate functions to present competing interpretations in isolation, because the disagreement itself contained valuable intelligence about where definitions, incentives, and assumptions were misaligned. Rather than treating those differences as a problem to suppress, I would use them to improve the organization's understanding of the opportunity. I would separate what was known from what was inferred and what was inferred from what people preferred to be true. I would also make the risks of both action and inaction visible, since organizations often examine the risks associated with pursuing an opportunity while failing to assess the cost of delay, the value of the options being surrendered, or the consequences of remaining with the current position. I would also spend more time understanding the chief executive's argument before challenging it, because people rarely change their position simply because more data is placed in front of them. Evidence becomes more persuasive when it is connected to the outcomes they care about, the risks they are attempting to manage, and the commitments they have already made. Challenging a decision effectively is not about overwhelming a leader with additional information. It is about helping them see how the evidence changes the strategic choice in front of them. Most importantly, I would treat the situation as a leadership and organizational design problem, rather than as a contest between good evidence and poor judgment. The objective should never be to prove the chief executive wrong. It should be to help the organization make the strongest decision available, because strategic intelligence is not about winning arguments. It is about improving outcomes. The opportunity the organization chose not to pursue cannot be recovered, but the lesson continued to shape my work because it showed me that organizations can have excellent people, extensive data, and advanced technology and still fail when the decision system surrounding those assets is weak. It also showed me that better outcomes become more likely when evidence, expertise, and experience are brought together early, when constructive challenge is treated as a contribution rather than a threat, and when leaders remain willing to revise a position before hierarchy, certainty or organizational momentum closes the discussion. Since then, I have focused increasingly on helping organizations develop strategic intelligence as a capability at every level, rather than treating intelligence as something produced by a specialist team and presented to executives at the end of a process. This means improving how people recognize meaningful signals, interpret evidence, challenge assumptions, design decisions, execute choices, and learn from outcomes, while ensuring that leaders create an environment in which uncomfortable information can travel without being softened, delayed, or ignored. The experience also changed how I understand leadership. Leaders do not simply receive the intelligence their organizations produce. Through their questions, reactions, incentives, and willingness to reconsider a position, they shape both the quality of the intelligence that reaches them and the honesty with which it is communicated. When leaders punish unwelcome evidence, the organization learns to hide it. But when they test their own assumptions and respond constructively to challenge, people become more willing to surface risks, identify opportunities, and contribute what they genuinely know. That is what strategic intelligence is designed to achieve. It is not about removing judgment, replacing leadership with data, or delaying action until certainty becomes available. It is about making judgment more disciplined, leadership more responsive to reality, and the organization more capable of learning before the cost of being wrong becomes too high. The decision I witnessed was made before the evidence was genuinely considered, and much of my work since has focused on helping organizations ensure that their most important decisions are shaped by evidence, experience, and expertise before positions harden and meaningful options disappear. This concludes the article.

Closing Thoughts And LinkedIn Read

SPEAKER_00 24:27

You can also read this article on my LinkedIn page where I share regular insights on AI, strategy, and emerging technologies.