Interpretating weak signs of business intelligence is an ill-structured problem encountered by many companies' managers. Those managers do not know how to interpret weak signs in order to extract meaning from anticipatory and fragmentary information elements. Surveys and action researches performed by our team confirm this problem and therefore, it becomes our main research focus. Because there is little underlying theory available on the subject, the design of a system helps to collect the reactions of end-users and to pick up relevant observations with a view to increase the understanding both weak sign interpretation and the business intelligence process. The method, the system and the issues that emerge from their validation are described.