Personal computers with Internet access are the most common way to transfer information nowadays. Furthermore, messaging technologies give us the ability to create dynamic interactions between systems. The purpose of this research work is to introduce an interactive messaging system that allows for a human user to communicate with a dedicated software application via messaging dialogues. The conversation is modelled as a system of interconnected state machines. Several techniques are presented that allow to manipulate complex dialogues preserving integrity, even in multiple languages. A smoking cessation case study is also presented, in which the proposed system holds 6-month long personalized conversations with all users concurrently. In such case study, the system monitors the smoking habits of these users independently and helps them quit smoking. The two-tier architecture provides a flexibility for relatively quick system updates when intervention protocols are revised.