An agentic framework to achieve systematic facilitation in group consensus building

Consensus building is a core challenge in human-centered systems, particularly in group decision-making contexts. However, effective consensus formation remains difficult due to social dynamics, cognitive biases, communication barriers, and asymmetries in information and power. Existing approaches offer partial solutions but suffer from key limitations such as over-simplification of features, limited adaptability to dynamics, and lack of systematic overall organization. To address these issues, we propose an agentic consensus building framework (ACBF) in which AI agents act as facilitators or collaborators to support human participants throughout the online decision-making process. This framework integrates participant modeling, consensus process management, and intelligent facilitation. Agents define simple tasks, decompose complex consensus challenges into tractable subtasks, and orchestrate their execution to ensure coherent and goal-aligned outcomes. This work contributes to the design of next-generation human-centered group decision-making systems that integrate agentic AI to systematically support online collaborative consensus building at scale.