The development of conversational AI technologies was mostly in the domain of engineers and hasn’t yet offered opportunities for non-technical experts to be engaged in the design process. To create a human-centric service experience with conversational agents, it is important to discuss the values, roles, needs of the users, and to embed these insights. In this paper, we aim to address the approach to designing a conversational agent with specific focus on identifying ‘roles’ of conversational agents to deliver customized services to its users. We propose a Role-Based design framework and tools, introducing 9 archetypal roles/purposes of conversational agents, resulting from case study research of conversational interfaces available in the market. The first part of the framework concerns the definition of the role of CAs, with respect to its intended purpose. Defining the role of a CA is key in order to design the user experience since it determines the expectations of the users in terms of performance, behavior, and service delivered.