AI Don't Fix Stupidity Game

You have been handed the AI programme. Eighteen months, a sceptical board, a budget that was set before anyone understood the problem, and an organisation that has already survived three transformations it did not want. Good luck. WHAT THIS IS A management game about doing AI transformation properly, built on the same method as the book "AI Don’t Fix Stupidity". You make the calls a real programme lead makes: what to fund, who to back, what to kill, and when to tell the board something they will not enjoy hearing. WHAT YOU ACTUALLY DO • Choose which initiatives to fund, and which promising-looking ones to stop before they eat the year. • Build a coalition. The CFO, the CIO, the sceptical operations director and the enthusiastic-but-unfocused innovation team all want different things, and you cannot have all of them on side at once. • Handle incidents. Models drift, a vendor misses a date, someone ships something that should have gone through review, and the story reaches the board before you do. • Watch the numbers that matter, delivery, adoption and trust, rather than the ones that look good on a slide. WHY IT IS HARDER THAN IT LOOKS Every option is defensible. The wrong move is rarely the obviously stupid one; it is the reasonable one taken a quarter too early, or the good idea backed without the coalition to survive its first setback. Most runs stall for reasons the player can see coming in hindsight, which is the entire point. BASED ON REAL PROGRAMMES The scenarios come from the method in Magnus Oxenwaldt’s book: the Six Shifts, the Earned Autonomy Ladder, the Tool Risk Ladder and the AI Heartbeat KPI. If you have read it, you will recognise the shape of every decision. If you have not, you will finish a run with a fairly good idea of why these things fail. PLAY IT IN SHORT SESSIONS A run fits into a commute. Progress is saved as you go.