“Practice makes perfect.” In the pursuit of honing our negotiation abilities, engaging in simulated negotiations is invaluable. These rehearsals not only allow us to grasp the nuances of negotiation principles but also may improve the negotiation result in the real negotiation. Reflecting on my experience with the EASE exercise, it became evident that deliberate practice and informed evaluation significantly contribute to skill enhancement. However, finding a suitable practice partner and a seasoned negotiator well-versed in all negotiation skills, capable of providing real-time feedback, can be challenging. This paper proposes an innovative solution — a Large Language Model (LLM)-driven chatbot designed to simulate negotiation processes for general, non-legal issues. The chatbot aims to provide users with a practical and interactive platform for refining their negotiation abilities.
The proposed chatbot will incorporate principles of negotiation and cover negotiation aspects such as process, relationship, and substance. Users will be able to engage in simulated negotiations with the chatbot by providing some basic information about the negotiation in format of negotiation preparation form we used in the workshop. The negotiation chatbot will be randomly assigned different personalities and employ diverse difficult conversation tactics. To enhance the learning experience, the chatbot will also offer real-time feedback and suggestions based on meticulously prompt-engineered negotiation knowledge.
This project is based on the Negotiation workshop in J-term at Harvard in Jan. 2024. Really appreciate the guidance from Professor. Dr. Alain Verbeke