I work with AI every day and, precisely for that reason, try not to give it authority it does not have. Using it does not make a process innovative: the difference lies in how initiative, verification and responsibility are shared.
In my practice, AI explores alternatives, speeds up systematic tasks and helps examine structures. The final decision still needs context, sensitivity and a person who takes responsibility for its consequences.
Documenting the collaboration
Recording only the name of a tool says very little. I find it more useful to explain what information it received, what it produced, why I chose an option and which limits I encountered.
Keeping that record helps me revisit decisions and learn from mistakes, including when an AI response initially seemed convincing.
The aim is not to automate everything
Sometimes I ask AI to open up alternatives; at other times, to examine a specific problem while I keep exploring. I adapt that division of work to what I need to resolve.
Working this way requires being able to disagree, revisit earlier work and stop an action when it is not sufficiently clear.
Explore Coordination Hub’s human–AI protocol