Abstract
The PRESTO (Plausible Representation of Emergency Scenarios for Training Operations) project, involving industrial and research partners, aims at adding semantics to a virtual environment and modularising the artificial intelligence controlling the behaviours of NPCs to support a productive end-user development environment. This report provides an overview of the approach being followed and a few preliminary results at the end of the first year of work, concerning in particular knowledge representation, the agent framework being developed and its graphical programming environment.