Recent developments in games and interactive storytelling applications have seen artificially intelligent computer controlled characters being included extensively. Non-human controlled characters are starting to play an increasingly significant role in enhancing the perceived intelligence of games. Although many of them employed certain cheating techniques (e.g. allocating more resources at the start to AI opponents to make them appear more aggressive), some limited learning did appear in several games (e.g. letting AI opponents remember where human users initiated attacked in previous game). In our Virtual Singapura research project, we incorporate software agents into our virtual world to provide more complex user interactions. With intelligent software agents being infused into interactive digital media applications, there is great potential in improving the overall user experience. However, during the process of our research, we discovered that the traditional way of adding a multi-agent system into a computer game requires a large amount of investment in time and resources and a high level of expertise in Agent Oriented Software Engineering (AOSE). Moreover, game AI is usually closely coupled with other parts of the game code which makes it hard to reuse or replace. This research proposes a multi-agent development and runtime framework which not only provides ease-of-use agent design and implementation tools but also can be easily plugged into various interactive digital media applications.