Dynamic Re-Configuration of Conversationally Initiated Automated Negotiations

Autonomous agents are well established to assist customers in contract negotiations on electronic markets. Recently, AI-based conversational agents provide more natural customer communication. Yet, on markets with varying negotiation items natural conversation can not be guaranteed to include all possible customer preferences for optimum negotiation results. In this paper, we propose a negotiation model supporting dynamic interactive preference post-specification and present a Hybrid Alternating Offers Protocol for that purpose. We demostrate it for conversationally initiated negotiations in the context of an app-market.