Intelligent information processing systems, including search, question answering, and summarization systems, must be able to handle pronouns, i.e. elements that do not have independent reference. We thus, focus on pronominal anaphora resolution (Mitkov [20], Kennedy and Boguraev [18], Lappin and Leass [19]), and identify the main features of an intelligent software design that includes the identification of minimal domains of interpretation based on asymmetric relations. We show that the identification of domains of interpretation based on asymmetric agreement (Di Sciullo [9]) opens new avenues in intelligent systems design.