The judgment AI does not replace
Why capable teams with good tools still make worse decisions than they did before.
What it covers
AI can produce the work. Someone still has to set the objective, direct the work, check the output, and own the outcome. When those four responsibilities are unassigned, a fluent and plausible answer travels through an organisation with nobody behind it. This session names the four, shows where they break in practice, and gives leaders a way to test their own decisions against them before they leave the room.
What the audience leaves with
- The four responsibilities that do not transfer to a model, in plain language
- How a confident wrong answer acquires authority on its way up a reporting line
- A test to run on one recurring decision in your own organisation next week
- What to stop putting in an AI policy, because it belongs in a job description
Book a date
Put someone on stage who measures what he is talking about.
Tell me the date, the city and the audience, and I will tell you honestly whether I am the right person for your programme. Next confirmed: 10th TPO Forum, Taichung, Taiwan, 16 to 18 September 2026.