The current organization of Consumer Price survey in Italy is based on a nonintegrated approach where data is collected either on paper or on portable PCs and handled by a software written in COBOL. Data transmission to Istat is performed by loading sequential files on a web site. After the Municipal Offices of Statistics (MOS) have loaded micro data, Istat recovers them on its own archives and after a detailed data check, edits micro data. The current process is not integrated because Istat archives and MOS archives are physically separated entities: therefore intervention carried out on Istat archives has to be repeated on the MOS archives. It is evident that the use, though partial, of paper questionnaires and the archives duplication are sources of non sample errors due in particular to measurement errors. This makes it more and more difficult to accomplish in an efficient way the increasing quality requirements in terms of accuracy coupled with timeliness and coverage. A project currently being carried out at Istat has the aim of realizing a in-depth reengineering of the whole Consumer Price (CP) survey IT environment, in order to reduce the possible sources of non sampling errors and improve the possibility to measure data quality through different set of indicators. The new IT architecture is based on a centralized relational database that stores all the survey data. The new data collection application, running on mobile PCs, allows data collectors to load data directly into the database right after they are collected. The new control and correction application allows to perform checks and editing on micro data directly on the database, with no confusing data redundancy. Moreover it makes available different set of indicators to monitor data quality. In this paper we describe in detail the new IT architecture that will support the survey and the improvements in data quality expected from its introduction.