Insights
The questions Andre hears most often after a talk, answered in plain language and without the sales pitch.
In eighteen years of building software with artificial intelligence, the question that follows the work has never changed: should people be afraid of this? The honest answer is that fear is a poor guide. It hides the specific decisions leaders, teachers and pastors actually face, and it gives the people selling tools an advantage they do not deserve.
What is true is that the tools are useful now, and will be more useful next year. They already draft, summarise, search and translate well enough to change how a small team spends its week. They are still poor at knowing what matters, admitting uncertainty, and carrying responsibility for the result. Those three gaps are where people stay essential, and where the organisations Andre talks with find their advantage.
That depends far less on the technology than on the tasks you do. Work built on repetition and pattern-matching is being absorbed quickly. Work built on trust, judgement and accountability is not, and it is growing in value. The practical move is to learn which parts of your job AI now does well, hand those over, and get better at the parts only a person can do.
Pick one repetitive task that already has a clear definition of done, and try it for two weeks. Keep the process change small enough that you can measure it. Most failed AI projects are not technology failures; they are projects that tried to change three things at once and could not say afterwards whether anything improved.
Yes, with disclosure and care. Using it for scheduling, first drafts and research is ordinary tool use. Passing off generated text as your own work, or feeding personal information into systems you do not understand, is where organisations get hurt. The rule that covers most cases: use it openly, and never let it stand in for a relationship you owe someone.