This paper is part of a research project called “A Computational Model of Distance Learning based on the Socio-Interactionist Approach”. This project is related to situated learning, i.e., in the conception of cognition as a social practice based on the use of language, symbols and signs. The objective is the construction of a Distance Learning environment, implemented as a multi-agent system composed of artificial and human agents, and inspired by Vygotsky’s socio-interactionist theory. This paper aims at the conception of two of the agents from such architecture: the Semiotic and the Collaboration Agents. The Semiotic Agent is responsible for searching adequate instructional material in the database to be presented to the student. The Collaboration Agent is responsible for assisting the interaction among students in a collaborative communication tool and it will consider the cognitive, social and affective capabilities of the students, which becomes a more qualitative mechanism for learning.