The contribution of intelligent agents for the human team performance is a challenging problem. This paper introduces a study to help clarifying this issue, starting with the motivation behind mixing agent teams with human teams. The main research problem is defined, as well as the key hypotheses. The validation methodology based on Psychology studies requires the definition of team and individual performance assessment measures. These measures will be the basic input to create a teamwork model able to help estimating team performance. The research aims to implement and test this model in a multi-agent environment to support the geographically dispersed human teams on trial cases, both at team and individual levels. This paper content is qualified as work in progress. KEYWORDS Teamwork Team Performance Agents