Leaders are paid to make high-stakes decisions with incomplete information. Every single day, they decide which numbers to trust, which experts to believe, which forecasts to act on, and which explanations deserve another question.

The uncomfortable part is that our brains are not evaluating these things on evidence alone. We are also responding to built-in psychological cues: confidence, familiarity, presentation, authority, and narrative. Which means a convincing answer can win before the correct answer ever gets a chance to enter the conversation.

This is why understanding what makes something feel true is not a philosophical exercise. It is a business skill.

We’d like to think we are making decisions by weighing the facts in front of us, but humans are constantly using shortcuts to determine what deserves attention and belief. Usually, that’s helpful. If we had to independently verify every sentence we heard before moving on with our day, none of us would get anything done. The problem is that those shortcuts can be manipulated, or simply misfire.

A confident (or just…loud) person can sound more credible than a cautious one. A statistic repeated often enough starts to feel familiar. A beautifully designed chart can create a sense of objectivity that the underlying analysis may not deserve. A story with a clean beginning, middle and end is easier to believe than an explanation that admits reality is messy.

None of those things inherently make the presented information true. They make it easier to accept. That distinction can be powerful if recognized, and costly if overlooked.

Put on your behavioral psychology hat for a second. A lot of what we call “credibility” is really a collection of signals. We notice whether someone speaks with confidence. We notice whether the information is presented cleanly. We notice whether we have heard something before, whether it fits with what we already believe, and whether the source has status or authority.

Those signals aren’t useless. Expertise matters. Presentation matters. Repetition can reflect consensus. But they are still signals, not proof.

There is also an awkward inverse: true information can feel less credible when it is uncertain, complicated, or poorly communicated. The person who says, “I’m about 70% confident but with two important caveats,” may be more accurate than the person who says, “This is definitely what’s happening.” Unfortunately, certainty is much easier to process than nuance.

In business, that can create a strange incentive structure. The person with the cleanest story can end up carrying more influence than the person with the strongest evidence.

One reason repetition works so well is that familiar information takes less effort to process. If you’ve heard the same claim enough times, it begins to feel natural. Not necessarily because you’ve seen better evidence, but because your brain no longer has to work very hard to understand it.

Presentations can have the same effect. A polished executive deck with clean charts and confident language feels more authoritative than a messy spreadsheet full of caveats. Sometimes that’s because the polished version is genuinely better conceived. But sometimes it’s because somebody knows how to use PowerPoint (or Claude) really well.

I say that with love 💚

The same thing happens with data. We tend to grant numbers a special kind of objectivity, especially once they are sitting inside a dashboard. But the number did not choose its own definition! Someone selected the time period, decided what to exclude, chose the denominator, set the chart scale, and determined which metric deserved the biggest font on the page. The data may be accurate and the presentation may still shape what feels true.

This gets more interesting with AI because fluency is persuasive. A good AI response is often clear, organized, confident, and remarkably fast. Those are all behavioral characteristics we as humans naturally associate with competence. The answer arrives fully formed, with none of the visible uncertainty or awkward pauses we expect from another human being trying to work something out. That can make it dangerously easy to confuse coherence with correctness.

This is not an argument that AI is uniquely deceptive. Humans have been confidently wrong since long before the chatbot arrived. The difference is scale. We can now generate persuasive-looking explanations faster than we can verify them, which makes human judgment more important, not less.

I love storytelling. It’s one of the best tools we have for making complicated information understandable. It’s also a very efficient way to make complicated information appear simpler than it is, however misleading that may be.

We naturally prefer explanations that show clean cause and effect. Sales fell because of this. The project failed because of that. Customers behaved this way because of one clear motivation. Real life is usually less cooperative.

Sometimes the honest answer is that five things happened at once, the data is incomplete, and we are not entirely sure which factor mattered most. That answer is intellectually responsible and narratively terrible. So we clean it up.

The danger is not storytelling itself. The danger is when the elegance of the story becomes evidence for the story.

I do not think the solution is to become suspicious of everything. That sounds pretty exhausting. The more useful habit is to notice when something feels convincing unusually quickly and pause long enough to ask why:

If the answer is “nothing”, I am probably no longer evaluating the evidence. I am defending a belief.

The goal is not to automatically distrust every chart, expert, executive or AI response that crosses your desk. The goal is to get better at separating persuasion from proof. Because the most dangerous information is not always the obviously false kind. Sometimes it’s the information that feels true enough that we stop questioning it.

If your team is making important decisions from reports, AI outputs, forecasts, or dashboards that nobody has really pressure-tested, ThinkFig can help. I work with organizations to examine the logic underneath the answer, clarify what the data actually supports, and build decision systems people can trust for the right reasons.

-T