Due to the high availability of data, users are frequently overloaded with a huge amount of alternatives when they need to choose a particular item. This has motivated an increased interest in research on recommendation systems, which filter the options and provide users with suggestions about specific elements (e.g., movies, restaurants, hotels, news, etc.) that are estimated to be potentially relevant for the user. Recommendation systems are still an active area of research, and particularly in the last years the concept of context-aware recommendation systems has started to be popular, due to the interest of considering the context of the user in the recommendation process. In this paper, we describe our work-in-progress concerning pull-based recommendations (i.e., recommendations about certain types of items that are explicitly requested by the user). In particular, we focus on the problem of detecting the type of item the user is interested in. Due to its popularity, we consider a keyword-based user interface: the user types a few keywords and the system must determine what the user is searching for. Whereas there is extensive work in the field of keyword-based search, which is still a very active research area, keyword searching has not been applied so far in most recommendation contexts.