BUILDING A COMMUNICATIONS TEAM THAT IS NOT HUMAN

BUILDING A COMMUNICATIONS TEAM THAT IS NOT HUMAN

The white paper “Building a Communications Team That Is Not Human” by Nanne H.C. Bos, published by the Scriptorium Initiative, documents an experiment on the question of whether the internal dynamics of a communications team can be replicated with agentic AI. The focus is on reconstructing the decision-making processes that shape communication before visible results emerge.

The shift in perspective is central: communication teams are not seen as mere production facilities; they form systems for decision-making under uncertainty. Quality therefore arises from structured debate. Different roles such as Project Leads, Stakeholder Intelligence or Communications Strategist specifically introduce tensions that are not resolved but made productive. The decisive factor is therefore not so much consensus as the way in which dissent is organized. Agreeing too quickly leads to interchangeable results, unfiltered divergence to inconsistency. High-performing teams move between the two, supported by shared understanding and clear roles.

Building on this, the paper identifies a gap in current AI architectures. In addition to tasks (what is done) and context (what is known), an explicit intent level is often missing. It defines which goals are prioritized and how goal conflicts are to be resolved. In the prototype, this level is operationalized by several specialized AI roles interacting with each other and negotiating decisions. The guiding principle here is an overarching one: communication primarily serves to build trust. Metrics such as reach or efficiency remain subordinate. The design of intent thus becomes the central control task in agent-based systems.