.svg)
.png)
Not all friction is waste. Some of it is where the thinking happens.
AI is very good at stripping out real inefficiency, the scheduling back-and-forth, the routine email, the meeting summary nobody wanted to write. Good riddance to all of it. But the same tools that clear away pointless friction make it tempting to clear away the useful kind too: the pause where someone notices the missing customer segment, the risk buried in a footnote, or the fact that the clean solution solves the wrong problem.
This gets harder exactly when AI output looks best. People treat information that's easy to process as more credible, and AI can produce fluent, confident, structured answers that are quietly wrong. When something looks finished, the instinct to question it switches off.
The fix isn't distrust. It's discipline: knowing when to use AI, when to check it, and when to keep it out of the work entirely. And it means deliberately placing a pause where work is about to travel, before a recommendation leaves the team, before a forecast lands in a board deck.
None of this requires a committee. It requires one thing: making it safe for someone to say "this looks wrong", and treating that as part of the work, not an interruption to it.
This article is an extract of "The Human Skills Premium: Where Value Moves in the Age of AI" by Simon Blanchette, McGill University Desautels.
