The increasing percentage of elderly people in the population (at least in Europe and North America) is part of a demographic change that will have an enormous impact on the society in the next few years. Thus, intelligent solutions that rely on the Internet of Things have been proposed in the literature, with the final goal to give remote support to elderly people at their home. Those solutions aim at monitoring activities and behaviors, and automatically send alarms in case of anomalies, putting in contact the end-user with her/his GP or alerting the emergency center or familiars, according to the specific needs. However, although the great deal of worldwide researches in the area of the Internet of Things and its early applications to healthcare and teleassistance, the technologies to apply it in real-world with the necessary dependability levels are far from being assessed. In this paper, we propose novel software engineering concepts that, by synthesizing existing proposals bringing in the lessons of agent-based computing and agent-oriented software engineering, can effectively support the systematic (and thus more dependable) development of Internet of Things applications.