I have done it myself. Let AI take the wheel, follow it down three tangents, and surface with a pile of output that was technically responsive … but entirely beside the point. The volume went up. The value did not.
Our teams are busy. Productive, even. Decks are getting built, analyses are landing, and recommendations are moving up the chain. All of it is polished.
The problem is never that the output looks bad. It is that it looks so good nobody thinks to question it. Polished-but-wrong does not just waste time. It erodes the trust that transformation depends on.
But there is a deeper problem underneath that one.
AI has no inherent sense of direction. It will generate a brilliant answer to the wrong question with exactly the same confidence it brings to the right one. Which means the question is no longer just “are we doing this well?” It is “why are we doing this at all, and who does it actually serve?”
That is what discernment is. Not just the ability to judge quality, but the ability to judge purpose. To stop and ask whether the work in front of us connects not only to the outcome we are accountable for, but to the impact we are actually trying to have in the world.
Speed and volume used to be competitive advantages. AI commoditized both. What is scarce now is the leader who can look at confident, coherent output and ask whether it not only moves the needle, but serves its purpose.
Most organizations treat discernment as a personality trait you either have or do not have. It is not. It is a capability you build deliberately.
Building it requires something most organizational cultures actively work against: the willingness to slow down when everything is pushing you to go faster. To sit in the discomfort of a question longer than feels productive. To say “I don’t think we’re solving the right problem” when the deck is polished and the meeting is almost over. That is not obstruction. That is leadership.
In practice, it looks like this:
Ground decisions in purpose, not just quality. Not just “what is the best answer?” but “what are we actually trying to move, and does this make things better for our customers, our people, the business?” That question, asked consistently before output gets acted on, is the single most important habit shift.
Make purpose operational, not aspirational. If the why only shows up in planning decks and all-hands presentations, it will not influence decisions. Leaders have to actively connect the work in front of them to the impact they are accountable for in everyday conversations, not just quarterly reviews.
Reward thoughtful dissent, especially when it challenges direction. Discernment requires people to say “this doesn’t get us where we need to go” even when the data says go and the deck looks great. That only happens if senior leaders model it and recognize it when they see it.
Teach your team to interrogate AI, not just use it. The question is never just “is this correct?” It is “is this right for what we are actually trying to accomplish, and for whom?” Every leader using AI needs to ask whether the output connects to something that genuinely matters and push back when it does not.
Build reflection into execution. After significant decisions, ask: did this move us closer to the impact we were trying to have? That loop is how discernment gets stronger over time. Skip it, and you are optimizing for output volume rather than judgment quality.
Discernment is a muscle. Purpose is what trains it. If you are not working both deliberately, you drift faster and faster, in the wrong direction.
We have been handed an extraordinary tool. The organizations that use it well will not just be the ones with better processes for reviewing AI output. They will be the ones that stayed clear on why they were using it in the first place.
