For most of our careers, the quality of someone’s work has been a useful, albeit imperfect, signal of their capability. Someone who consistently produced strong analysis probably knew how to analyze. Someone who repeatedly articulated strategy in a compelling way was probably developing executive-level judgment.

The questions we hear executives asking are good ones. “How should AI change our business model?” “Which processes could an agent run?” “Where can we reduce costs?” Those are the right questions to be asking, and most of the companies we work with are starting to get real answers.

The question we hear far less often is, “What might these efficiency gains cost us in the long run?”

Here’s the caution: leaders can capture near-term efficiencies that improve the business today, while unintentionally disinvesting in the growth of the people who will differentiate it tomorrow.

Leaders would never make this tradeoff on purpose. It happens because the efficiency gain is visible immediately and missed talent development stays hidden for years.

The work can improve faster than the person

A person can now produce work considerably more sophisticated than their experience would have allowed a few years ago. That is a real productivity gain, and we should want it. However, the work can improve much faster than the person producing it.

Let’s say two people have the same laptop setup, the same assignment, and the same hour to complete it. One asks AI what to think and then edits what AI provides. The other develops a point of view and then brings in AI to attack it, look for what they missed, and change their thinking when the evidence warrants it. Both people use AI. Both may produce acceptable work. From the outside, the two workflows look nearly identical. In the short term, these efficiencies appear harmless.

Man in a blazer and glasses working on a laptop at a table by a window

Intellectual tourism

Whether we realize it or not, the silent undercurrent of the AI talent gap we’re headed toward is an outsourcing of thinking. In his TED talk, How to Stop AI from Killing Your Critical Thinking, AI & Design Researcher Advait Sarkar calls this concept “Intellectual Tourism.” We visit ideas, but we no longer inhabit them with the rigor to understand, defend, and embody our point of view. He argues that we have fewer ideas, think about the ideas we do have less critically, and remember them less well.

A team at the MIT Media Lab offered an early glimpse of this in 2025. Researchers put EEG caps on 54 participants and had them write essays over several sessions. One group used ChatGPT, another a search engine, and a third used neither. The researchers found significant differences in the level of brain connectivity across the groups. More striking was what happened after the writing was finished. Minutes later, 83 percent of the ChatGPT group could not accurately quote a sentence from the essay they had just produced. Comparatively, in the other two groups, only 11 percent struggled to remember what they just wrote.

This is an early, small study, and we wouldn’t build a whole talent strategy on it. However, both this study and Sarkar highlight the risks of people producing work without thinking it through.

Man standing and smiling at a whiteboard marked AI while a small team sits around a table

Where judgment actually comes from

The people on your team today need to develop the judgment, insight, and perspective required to run your company in the future. The successful leaders of the future will exercise and hone the craft of critical discernment.

Historically, much of that development happened through the work itself. You got an assignment slightly beyond your ability. You wrestled with it. You made decisions. Somebody more experienced looked at your work and told you what you missed. Then you did it again. Enough repetitions eventually turned into well-earned judgment. Useful wisdom.

AI changes that equation because it lets us remove friction from knowledge work at extraordinary speeds.

Two kinds of friction

The first kind of friction was wasted effort and should disappear. Nobody needs an employee spending hours formatting a document to build character. Nobody becomes a stronger executive by manually performing work that software does accurately in seconds.

The second kind plays a key role in a person’s development. Forming a hypothesis before seeing a polished answer. Deciding which evidence matters. Making a recommendation before knowing whether the boss agrees. Being wrong, learning from it, and trying again.

Organizations have rarely had to distinguish between the two kinds of friction because both were built into the work. Now, they do. Remove both kinds of friction indiscriminately, and you make today’s employees considerably more productive while quietly removing the experiences that would have made them tomorrow’s leaders. This is the talent gap.

Friction to remove:

Formatting the document
Retrieving information
Repetitive first drafts
Clerical reconciliation
Summarizing what was said

Friction to protect:

Forming a hypothesis before seeing the answer
Deciding which evidence matters
Recommending before the boss weighs in
Reconciling conflicting information
Being wrong, hearing about it, and trying again

Coaching gets skipped

Most organizations already struggle to get enough useful coaching conversations between managers and employees. AI makes avoiding those conversations easier. Imagine an employee produces something that looks good, but is too generic to be useful or isn’t likely to work in the context it’s designed for. A manager can fix it themselves, ask AI to improve it, or decide it is good enough and move on. All of those options except coaching are faster. So the manager becomes more productive, the employee produces more work, the organization sees increased efficiencies throughout, and nobody gets better. A company can look increasingly efficient while the development of its people quietly erodes.

Imagine if regular coaching conversations focused on the ownership of the thinking. Can this person explain how they arrived at a recommendation? Can they name the assumptions carrying it? Can they say where AI influenced the work and where they disagreed with it? Can they tell you what would have to be true for them to change their mind? Those questions tell us more about someone’s readiness to lead than the output of their work alone.

We have been living with these questions inside Sagency. Our team has heavily invested in powerful AI tools and uses them regularly. We, too, value intentional efficiencies, just as the clients we serve do. Paired with that desire for efficiency is a shared expectation: own the work, and be able to defend the thinking behind it.

The choice in front of you

Whether your people grow or escape to intellectual tourism over the next three years is a choice your organization is making right now, mostly without noticing. Three years from now, someone on your team will need to step up and lead. What will your organization do today to redesign work so they’re prepared for it?

The authors of this article, Mike Meagher and Ashley Blazek, are both part of the team at Sagency. Sagency is a strategic advisory, leadership development, and executive search firm that helps organizations get stronger as they grow through Strategic Clarity, Leadership Capacity, Talent Depth, Execution Discipline, and High-Trust, High-Ownership Culture. Learn more at sagencytalent.com.

Portrait of Mike Meagher, smiling, wearing a blue plaid blazer

Portrait of Ashley Blazek, smiling, wearing glasses and a blue top, seated on a cream sofa

Learn more: Sagency | Sagency on LinkedIn | Sagency on Facebook | Mike Meagher on LinkedIn

Sagency

The Hidden Costs Of AI