Your Best Leader Can Become the Biggest Blind Spot in Your AI Transformation

The leader most likely to fail at an AI transformation is not the careless or incurious one. It’s the confident, accomplished, deeply knowledgeable one. That runs against every instinct about who you want in charge of a hard change, which is exactly why it’s worth sitting with.

Here is the evidence that something is off. In one recent survey, 85 percent of operations leaders said they were ahead of their competitors on digital transformation, while 89 percent of the same group conceded their technology investments had not fully delivered [1]. The confidence and the shortfall belong to the same people at the same time. Set that against the fact that 95 percent of organizations report no measurable return on their generative AI spend [2], and a pattern appears.

Capable leaders often believe they are on track while the results say otherwise. Feeling on track is not evidence of being on track. It is often the first symptom that something has gone wrong.

Why your most capable leader may be the biggest transformation risk

Because expertise, in new and unfamiliar terrain, quietly turns from an asset into a filter.

You spent decades building a refined, repeatedly validated sense of how value gets created in your industry. You were right often enough that the confidence hardened into identity. That is not arrogance. It is the honest result of evidence and years of positive feedback. The problem is that knowledge and wisdom are not the same thing. Knowledge is the pattern library you built from a particular context. Wisdom is knowing which of those patterns still apply when the context has quietly changed underneath you.

The old environment rewarded accumulated knowledge because the rules rarely changed. AI changes the rules, and the instrument that read the old conditions perfectly keeps reading them as if nothing moved.

The new situation looks enough like the old one to fool the instrument. Information that fits your existing model gets through. Information that doesn’t gets dismissed as hype, or as advice from someone who doesn’t understand your business. You are not ignoring the signal. You simply cannot see it, because you are using a proven map for uncharted water.

When confidence stops being a leadership strength

It is, right up until it becomes the thing that prevents you from noticing you’re wrong.

Dunning and Kruger are usually misremembered as having shown that incompetent people overestimate themselves. Their actual finding is more uncomfortable and more relevant here: the same knowledge gap that produces an error also removes your ability to detect it [3]. The blindness is built into the mistake. Applied to a transformation, this means the leaders heading toward the rocks and the leaders actually on course often report the same internal experience. Confidence cannot tell them apart. Only honest information can, and honest information is the first thing a confident, successful culture stops producing.

The pattern is more common than most executives realize, and we watched it surface while deploying an Intelligent Digital Worker for a large regional organization. For weeks, the confident conclusion in the room was that the AI kept failing. That conclusion was the blind spot. What the project actually exposed was how much of the organization’s real operating logic lived only in its people’s heads, never in a system, a point I made in an earlier article.

During user testing, the project repeatedly appeared to be failing. Every time the AI hit a transaction it had not been taught, or a situation outside the documented process, the immediate and confident assumption was the same: the technology had made a mistake. Nobody paused to ask whether the assumption itself might be wrong. That reflex, trusting your own read of the situation over the evidence in front of you, is exactly how a capable team’s blind spot operates.

The evidence told a different story. As both teams investigated case by case, most of the perceived failures were not software defects at all. They were undocumented business scenarios, inconsistent practices, requirements that shifted mid-test, ERP constraints, and decisions experienced staff made on instinct but had never written down. The AI was not failing to understand the business. It was holding up a mirror to how the business actually worked, and the reflection was uncomfortable enough that the easier response was to blame the tool.

The turning point came when the CEO stepped in and changed the question. Instead of ’Why didn’t the AI handle this?’ the discussion became ’What kind of issue are we actually looking at?’ Every case was classified honestly: a technology defect, a new business scenario, an operational dependency, an ERP limit, or an enhancement request. That one shift moved the room from defending its assumptions to examining them. Priorities got clearer, decisions got faster, and the project regained momentum.

Nothing fundamental changed in the technology. What changed was the leadership mindset. The breakthrough was a leader willing to question his own organization’s assumptions instead of blaming the tool, which is the capability most confident leaders lack until something forces the question. AI can reveal what an organization doesn’t know about itself, but only its leaders decide whether that evidence drives change or becomes one more reason to keep things as they are. AI doesn’t just automate the work. It exposes how the work actually happens.

The tell is usually misread:

You can’t diagnose this blind spot from the inside

You mostly can’t, and that is the entire point.

Petriglieri’s research at INSEAD [4] suggests that leaders who navigate disruption successfully are not the ones with the most refined skills, but the ones whose sense of themselves can absorb continuous revision under pressure. Traditional executive selection rewards knowing, deciding, and projecting certainty. AI rewards the opposite posture, in which not-knowing and revising are normal modes of leadership rather than failures.

The trap is that you cannot diagnose this from the inside, because the same conditions that create it also silence the people who could point it out. Absence of conflict reads as alignment. It is more often fear wearing the mask of harmony.

This is why a pilot can succeed on a team’s energy and technical merit but stall the moment it needs real organizational sponsorship. Scaling it means changing the assumptions that made today’s operation successful, and the leader who would have to back that cannot yet see what the pilot is telling them. That is a harder problem than building the technology.

What separates the leaders who succeed at transformation

They do something that sounds small and is extraordinarily hard: they change their own mind, on the record, and then act on it.

The leaders who succeed are honest about where they actually are. They push decisions down to where the real information lives, because the person closest to the customer already knows something the executive layer does not. They protect the people who think differently instead of selecting them out. And they get the order right: culture and foundation first, technology second, never the reverse.

In The Transformation Gap [5], an insight paper I co-authored in June 2026, we mapped nine human conditions that decide whether a transformation holds, and leadership cognition sits at the top of the list because everything else is downstream of it. A leader who cannot revise their own thinking cannot sequence anything below it correctly, which is why the honest work starts with the person, not the platform.

The one question worth more than any readiness assessment

Before you approve the next strategy, platform, or transformation budget, answer this.

Can the leader who will own this transformation point to a specific, recent moment when they changed their own mind because someone below them or outside the company challenged them, and then acted on it? If the answer is yes, with a real example, you are probably closer to ready than you feared. If the answer is no, or “I’m not sure,” you have just found the actual transformation problem. It is not a development note to revisit. It is the thing itself, and every technology decision is downstream of it.

That question is uncomfortable to sit with, and the discomfort is the most useful data you will generate this quarter, because it is the signal your own organization may no longer be able to send you.

The most valuable first move is not another tool or another roadmap. It is an honest read, ideally from outside the building, of where you actually stand and what has to change before the technology can do anything at all. That is the thinking behind The Transformation Gap and the diagnostic work around it. Commission that read before you approve the next transformation dollar, and let what it finds set the direction.

Sources

1. PwC, 2026 Digital Trends in Operations Survey: 85% of operations leaders say they are ahead of competitors on digital transformation; 89% concede their technology investments have not fully delivered.
2. MIT NANDA, The GenAI Divide: State of AI in Business 2025: 95% of organizations report no measurable return on generative AI.
3. Kruger, J. and Dunning, D., “Unskilled and Unaware of It,” Journal of Personality and Social Psychology, 1999.
4. Gianpiero Petriglieri, research on leadership identity workspaces, INSEAD.
5. Lisa Hyde and Jahan Ali, The Transformation Gap, The Counsel, June 2026.