Abstract

Software agents are being increasingly applied to e-commerce activities, including commerce
negotiations. Agents can be used to conduct negotiation tasks on behalf of users. When users
delegate negotiation tasks to agents, information technology plays a role in determining social
affairs. The locus of control over social affairs partially shifts from human participants to
technology. When this negotiation approach is adopted, an important question arises: how will
users treat and assess their agents when they delegate negotiations to agents? It is challenging to
develop agents that are able to connect with users in meaningful ways. This thesis argues that users
will not treat their negotiating agents in the same manner as they treat classical computer-enabled
tools or aids, because of the autonomy of the agents. When assessing agents, users will be heavily
oriented towards their relationships with the agents. Drawing on several streams of literature, this
thesis proposes that the notion of control helps to characterize the relationships between users and
agents. Users’ experienced control will influence their assessments and adoption of their
negotiating agents. Users’ experienced control can connect to instrumental control, which is a set
of means that empowers the interaction between users and agents. An experiment was conducted
in order to test these propositions. The experiment results provide support for the propositions.