Creativity and innovation are fundamental for the progress of individuals and society, and recognised as the engine of sustainable and competitive economies. Thus, it is a challenge for companies and for designers and/or design teams to come up with products that are both new and adapted to future users. Designers or design teams have to be creative in order to imagine and conceive new products but it appears particularly complex (Bonnardel, 2012). From a cognitive point of view, a main characteristic of design activities is that designers' mental representations are initially incomplete and imprecise. Only by going through the problem-solving process itself can designers complete their mental representations. It is even more difficult in the context of prospective ergonomics, which is turned towards the creation of future products that have not been identified yet (Robert & Brangier, 2012), since designers have both to detect existing user needs, to anticipate future ones, and to inject creativity in the design solutions they propose. Moreover, in complex design situations, it is not only a single designer but a design team that has to perform creative activities. In these situations, the knowledge required for performing the design task is often distributed amongst different stakeholders who have different perspectives and backgrounds. For instance, design teams may consist of designers and specialists in ergonomics and, sometimes, end-users. However, the participation of this last kind of stakeholders is not always possible, due to pragmatic constraints (time, cost, availability of end-users, etc.). In addition, these collective activities may have positive effects on creative design activities, since people adopting different viewpoints or perspectives can propose new ideas, but also negative effects, such as production blocking due to an apprehension of evaluation or a social inhibition (Paulus & Nijstad, 2003). Therefore, our aim is to contribute to defining design conditions that can facilitate both creative and collaborative activities in design teams. Towards this end, we first present a specific user-centred design method that we find particularly promising: the persona method (see, for instance, Bornet & Brangier, 2013), which enables designers to focus on archetypal future users of the product or service to be designed. This method is thought to favour both empathy and creativity among designers, but only recently have some studies highlighted certain benefits and limitations of using the persona method (Bornet & Brangier, ibid.). Thus, our objective in the present study is to analyse the use and impact of a new kind of persona, which has the particular feature of being 'dynamic', by comparison with a classical situation based on a 'static' persona. 2. Classical persona method vs 'dynamic' persona method User-centred design methods can help designers understand users' expectations, needs and limitations, and take them into consideration during the design process. Among them, the persona method is based on the creation and use of fictional and personalized prototypes of future users, called 'personas', which are associated with concrete representations comprising both textual descriptions and photos. In specific cases, personas may be played by a team member, which may be dependent on his/her performance as actor. The specificity of our research is that we propose a new kind of persona method based on 'dynamic' personas interacting through a virtual enviroment. Moreover, our aim is to compare its effects with the ones of a classical persona method, based on static descriptions. In our study, all interactions between the different stakeholders occur through a collaborative virtual environment (more precisely, Second Life) and in a written communication modality (or chat). These interactions may be less subject to social inhibition, especially, due to the anonymity that is allowed by the use of avatars. Information elements provided in the two kinds of conditions (dynamic vs static) are strictly the same but, in the case of the 'dynamic' persona condition, participants interact with a 'dynamic' persona who is played by the experimenter instead of reading the written description of the 'static' persona. Indeed, the experimenter cannot suggest ideas and he/she intervenes in the discussion in order to express characteristics, needs, and points of view of the future user, similarly to information elements that are provided in the 'static' persona.