Delegating the Deal: How Managers Decide AI Autonomy and Scope in Negotiations

Abstract: Organizations increasingly deploy AI systems to negotiate on managers’ behalf, yet little is known about how managers decide when to delegate negotiation to AI, how much autonomy to grant, and which parts of the process to retain. We argue that delegating negotiation is not a binary choice but an ongoing managerial decision about autonomy and scope, shaped by concerns about performance, alignment, and the subjective value of negotiating. We study these decisions in a repeated negotiation setting where participants can negotiate themselves, delegate fully to an AI agent, or retain control while using AI support. This design allows us to observe how delegation patterns evolve across negotiations and how autonomy and scope choices relate to economic outcomes, subjective experience, and willingness to continue negotiating. By shifting attention from whether AI negotiates well to how managers decide to involve AI in negotiation, this work aims to clarify the behavioral foundations of human–AI collaboration in core organizational processes.

Keywords: Negotiation; Human–AI collaboration; AI Autonomy and Scope; Subjective value; Delegation decisions