This research presents the concept of a non-command interface for a smart room to automatically detect when people talk about a document and whether it is present or not, as a fundamental prerequisite for missing document provision that doesn't require explicit requests, avoiding distraction from the main discourse. A study on how observers judge document usage in meetings is presented as a baseline and the conceptual framework is briefly explained. Finally, an exploratory experiment is reported. These elements demonstrate the research feasibility and define the techniques needed to build the agent.