Representing an epistemic situation involving several agents depends very much on the modeling point of view one takes. In fact, the interpretation of a formalism relies quite a lot on the nature of this modeling point of view. Classically, in epistemic logic, the models built are supposed to represent the situation from an external and objective point of view. We call this modeling approach the external approach. In this talk, we study the modeling point of view of a particular agent involved in the situation with other agents. We propose a logical formalism based on epistemic logic that this agent can use to represent `for herself' the surrounding world. We call this modeling approach the internal approach. We set some formal connections between the internal approach and the external approach and we axiomatize the internal approach. Then we show how this new internal approach allows for a straightforward generalization of AGM belief revision theory to a multi-agent setting. We also investigate some multi-agent rationality postulates in the AGM style. Finally, we independently sketch a general formalism for the external approach.