Researchers build an LLM mediator for pre-negotiation preparation

Researchers have proposed an automated mediator that uses a structured LLM pipeline to support pre-mediation before human negotiation.

The system breaks preparation into modules for dialogue, preference prediction, response critique, and structured summarization. That modular design is meant to avoid the brittleness of a single monolithic prompt.

The work points to a practical use case for LLM agents: helping people prepare for complex negotiations where trained mediators are too expensive or unavailable.