Email communication has been widely used in managing customer queries, such as complaints and inquiries. The user is expected to respond the query properly. However, with an increasing number of query emails received every day, it seems likely that the inbox appears to have more unreplied queries and leads to overwhelmed and cluttered email management called email overload.

This thesis presents a development in the email overload issue on managing a reply task. The system, called a smart email client, helps in replying to an email by suggesting a list of replies gleaned from the emails replied to in the past. These suggested replies are ranked according to the level of similarity. The methodology follows the Case-Based Reasoning (CBR) approach to solve the problem by reusing previously written solutions from the past replies stored in the case base. Techniques from Natural Language Processing and Information Retrieval are utilised, particularly in lexical semantics, through using WordNet. Finally, an evaluation of the retrieval algorithm shows that the effectiveness and efficiency of the algorithm is influenced by the case feature selection and various text analysis techniques such as lexical analysis, stemming, stopwords removal, and synonym expansion.