While learners can acquire vocabulary through extensive reading (Pigada & Schmitt, 2006), research suggests that acquisition can be more effective when supplemented with targeted vocabulary activities (e.g., Paribakht & Wesche, 1997). Problems arise, however, in determining what vocabulary learners have acquired, and what items should be focused on in these vocabulary activities. The main purpose of the current study is to describe the development and trial implementation of an intelligent system which created individualized vocabulary activities for each learner based on hyperlinked words that were clicked on during reading written passages in the target language. In this preliminary study, 43 Japanese learners of English were required to read several mid-length passages on computer during class time. Content words were hyperlinked to a separate window which provided meanings, pronunciation, and example sentences. Whenever a word was clicked it was added to a database and activities were automatically created for each learner that could be completed outside of class time, on either a computer or a mobile phone. Learners were given a pre-test at the beginning and a post-test on completion of the semester. Logs were kept regarding clicking patterns during reading. There were two specific objectives in the study. Firstly, the ways in which learners clicked on words on the passage were investigated to see if there were any differences in how words deemed as "known" and "unknown" were looked up, and whether this linked to acquisition of unknown words. Secondly, lists of the words looked up by the learners were analysed to determine if it was possible to create individual profiles of learners' vocabulary knowledge. Data collected in the current study included the correlation between the words considered as unknown according to the pre-test and the words clicked on during reading, and the length of time spent looking at word descriptions, and the results of the vocabulary post-tests.