Conversational Structured Hybrid 3D Virtual Environments

Structured 3D Virtual Environments are 3D virtual spaces where some users' interactions are regulated by a subjacent Organisation Centered Multi Agent System (OCMAS) ---an Electronic Institution (EI). They are task-oriented hybrid systems, where staff (organisational) software agents support --based on activities' specification and current EI state--human users in their task achievement. The contribution of this paper is a conversational task-oriented structured 3D environment, where users interact with staff bots (i.e. the embodiment of staff agents) using natural language and, as a result, the communication is improved. With this aim, we extend the Artificial Intelligence Mark-up Language (AIML) with special tags to enable the complex flow of task-oriented conversations. They are characterised by different conversation' states --such as asked, responded, failed or confirmed--and by data types from the EI ontology. We evaluate the usability of our conversational proposal and compare it to a previous command-based interaction system. As expected, data analysis on users' skills --command-based novice vs expert profile-- suggests the co-existence both conversational and command-based user-agent interaction styles. But overall, results show a higher users' satisfaction with the conversational approach which, in average, also performs better in terms of efficiency, effectiveness and errors.