Every firm now has access to the same capable AI. What separates one firm from another is the thinking its people bring to AI’s output.

Over the past months, I went through research from BCG, McKinsey and Deloitte, and from INSEAD, Wharton, MIT and Carnegie Mellon, to understand what is happening to that thinking. One BCG finding set the tone. BCG asked senior leaders what AI is doing to the skills inside their organizations, and the symptom they cited most was people accepting AI output without challenging it.

The research points to four frictions to remove now, and one to add on purpose: the Five Frictions of Thinking.

1. The meeting trap

Every problem starts in a meeting. The key people have to find the same hour, and the first ideas take shape out loud, in front of everyone. INSEAD and Wharton tested how teams should start. People who thought alone first came up with more ideas, and better ones, than teams that started together. When AI hands everyone the same first draft, thinking alone is the step most likely to be skipped.

2. The loudest voice

The best idea is often in the room, and nobody hears it. The loudest voice sets the direction, so at best the work gets redone, and at worst the team heads the wrong way. MIT and Carnegie Mellon studied hundreds of people working in groups. The best groups were not the ones with the smartest people, but the ones where people took turns speaking more evenly. AI adds a new loud voice, because its answer arrives first and rarely sounds unsure.

3. The broken ladder

Your best people can only teach when they’re free, which is almost never. Most of us learned our craft the same way: we wrote the first draft, a senior marked it up, and that review was where the teaching happened. Today the junior waits three days for twenty minutes, and the senior comes off a client file to give it. AI is now taking the first drafts and early analysis too, the very work juniors learned from. McKinsey looked at what this means for the next generation, and the shift is already visible in entry-level roles.

4. The exit gap

Firms chase knowledge on the way out because they never captured it along the way. Everyone runs exit interviews, and nobody gives their best thinking with one foot out the door. An exit interview records what someone did, not how they reasoned. In a large survey, Deloitte asked Gen Zs and millennials, many of whom already lead teams, whether their team could keep performing if a key expert left tomorrow. A large portion weren’t confident their team would cope.

5. The deliberate pause

Not all friction is bad. The path of least resistance is to rubber-stamp AI output, accepting what sounds right instead of testing what is right. The deliberate pause is where slow is fast: between the first answer and the decision, each person thinks alone first and questions what they hear, including what AI produced.

Remove the friction between your people and their best thinking, and add the one friction that activates their judgment.

The white paper goes deeper

The paper shares the numbers behind each friction, maps where thinking gets lost inside a firm, shows which skills BCG found most at risk, and explains what it takes to remove each friction.