One useful experiment
A practical starting point for AI in delivery.
By Iurii Kostenko · Field notes
Start where work gets stuck.
Choose one repeated task that takes attention away from people and decisions.
Watch the work before choosing the tool. Ask where information is copied, context is lost, or teams wait for an update. A specific friction point makes a useful experiment possible.
Make it small. Make it visible.
Give the experiment a clear owner, a short time frame, and an observable result.
Keep what helps. Change the rest.
Review the effect on the work with the people who do it. Decide the next step together.