Use intuition as a diagnostic filter
Don’t accept or reject AI recommendations indiscriminately: use your intuition as a litmus test. If something doesn’t feel right, go find out why. Allow your gut feeling to trigger deeper investigation.
Evaluate explanations, not just outputs
Teach yourself and your team to ask, “Does this explanation make sense in our context?”, rather than, “Do I like this answer?”
Create safe spaces for disagreement with AI
Intuition often manifests as hesitation or discomfort. Normalize the idea that it’s okay to question AI outputs, especially when explanations don’t feel satisfying. This avoids both blind trust and reactive dismissal.
Document when intuition conflicts with AI and learn from it
Humans often get it right when they question AI with good cause. Encourage teams to log and analyze instances where intuitive disagreement with AI led to better decisions, and build those patterns into future system design or training. The “traffic light” model below summarizes some tactics for managing situations when humans and AI agree or disagree.
