Seasonal adjustment is a very demanding activity for National Statistical Institutes (NSIs). The required resources, in fact, involve high skilled statisticians and econometricians, efficient IT procedures and a continuous updating of seasonal adjustment methods and tools. In order to conduct the annual campaign, as suggested by the European guidelines on seasonal adjustment, and to assure coherence in the treatment of peculiar situations (such as crisis, short series, seasonal heteroskedasticity, etc.), NSIs generally work through a well-structured organisational system. A unit service deals with time series and seasonal adjustment at central level while several other local units seasonally adjust data routinely. When a problem arises the latter asks for the help of the former. Given this organisation and chosen a common seasonal adjustment method (usually the choice is between the ARIMA model based method and the moving average based method), there may be several and different ways to approach the seasonal adjustment, ranging from the simplest one where common procedures are run on the own domain using personal computers to the most complex one where a unique centralised system is used. The latter approach is implemented in the Italian National Institute of Statistics (Istat) through an informative system hereafter called SITIC. It is aimed at storing, seasonally adjusting and disseminating short term statistics coming from all business surveys. The aim of this paper is twofold. Firstly the main features of SITIC are described stressing the solutions implemented to cope with some methodological issues concerning the aggregation of chained linked indices and drawing attention to the benefits implied by the management of a centralized system. Secondly the project aimed at introducing JDemetra+ in SITIC is presented. Issues, costs and advantages are listed and the peculiarities of the project are highlighted. In fact the transition to JDemetra+ in SITIC requires neither the use of the graphical interface nor the use of batch procedures like JDemetra+ Cruncher, but it is based on the integration of Java modules avoiding or minimizing any changes in SITIC interface.