1️⃣ AI design quietly changes who gets to decide INSEAD separates AI that helps managers direct and monitor work from tools that let colleagues coordinate directly. One strengthens vertical control; the other can move coordination sideways.
💡 Why it matters Choosing an AI tool is also an org-design choice. Decide whether it adds managerial oversight or removes handoffs.
☕ Coffee talk If the agent routes the work and flags every exception, who still gets to overrule it?
2️⃣ Anxiety can turn a transformation into theatre A four-year study of a professional-services firm found that unmanaged anxiety can push leaders and staff into defensive routines: blame and activity that leave the status quo intact.
💡 Why it matters More change plans will not fix a response that protects people from discomfort. Leaders need to make room for the anxiety before asking teams to absorb another change.
☕ Coffee talk When the steering group starts pointing fingers, is the problem still the plan?
3️⃣ Run two tests before backing one Jay Parikh says he would rather test two competing approaches than forecast a winner. A six-case study found simultaneous experiments can preserve strategic variety and long-term survival, but slow early growth.
💡 Why it matters Parallel tests buy evidence, not certainty. Keep both small and agree what result will end or combine them.
☕ Coffee talk If both bets stay funded, what will force the choice?