Session Overview
As organizations adopt AI to accelerate delivery, automate routine work, and support decisions, a fundamental challenge remains: how do people work together effectively when a team is still learning to trust, communicate, and self-organize?
AI can generate plans, analyze information, summarize discussions, and suggest solutions. Effective collaboration, however, also depends on trust, psychological safety, open communication, constructive disagreement, and collective ownership. Introducing a tool, framework, or process does not automatically create these conditions.
I will share practitioner experience using the Mafia social-deduction game in a rapidly formed delivery organization of more than 100 people. The case explores how structured social interaction can make team behaviors visible and create an opportunity to examine communication, assumptions, ambiguity, and collective decision-making.
This is not a game tutorial or a workshop about game mechanics. The practitioner case study and its implications remain central. A time-boxed, 15-minute Mafia activity, facilitated by my co-presenter, will let participants experience the dynamics under discussion. We will then examine what the activity reveals about participation, trust, influence, assumptions, and collective decisions.
We will connect these observations to familiar Agile challenges: silent retrospectives, uneven participation, dependence on managers, and reluctance to challenge assumptions or take ownership. We will then consider why these human capabilities matter as AI becomes more involved in planning, execution, and decision support.
Participants will leave with a practical lens for observing team interaction and questions they can use to translate experiential learning into stronger Agile practices and more effective human–AI collaboration.
AI can accelerate the work. But teams still need to learn how to work together.
