TAMIC-P aims at providing a single natural language interface to heterogeneous databases containing information on pensions. Pensions are particularly relevant, due to the nature of the multiple, diverse and distributed data involved and the fact that pensions organizations are large and have their own internal information processing centres. The interface will be used by clerks during their interactions with customers making enquiries on their pension positions and rights. It will shield clerks from the complexities inherent in accessing a number of different database systems and allow them to concentrate on providing improved access to information and better customer service. Customers will be able to get a clear picture of their own information The user focus for the project is Italian and Austrian social security organizations. The interface will use a combination of natural language and graphical user elements.
Pensions is a complex area with which even experts have difficulty keeping up to date. One reason for this is that in Europe the rules covering pensions are frequently changed. This makes it difficult for the citizen to find out the relevant information about his pension and social security arrangements. Automatic information kiosks would be a realistic option for only a limited number of citizens and only for dealing with the most trivial queries. The average citizen will need the help of a clerk who is normally competent in only some areas, and will expect one visit to the social security office to solve a complex problem. Most clerks are only competent in a limited set of areas. A clerk must not only determine the exact nature of the problem, but also find the answers and make sure the citizen understands them.
Social insurance organisations have many tasks, including managing contributions, paying and managing pensions and other social benefits such as unemployment, family support and maternity leave. Each one of these is complex. This complexity tends to create specialised departments, resulting in the fragmentation of information. Customers have to visit several people to obtain information which, even if different, is all to do with social insurance. Each department has developed specific administrative and technical expertise which is not easy to impart to others. This situation is hampering the provision of single access points able to answer all questions.
In Austria, social insurance is organised on a distributed basis: by region, by insurance category such as for pensions and health benefits, and by population groups such as private employees, farmers and civil servants. This often means that several social insurance agencies need to be contacted. Insurance agencies want to offer an all-round service, and this is currently held back by the incompatibility of the information stored in different databases with different physical and logical structures.
A further complicating factor is that in Europe, the trend is toward establishing communications networks between different administrations and with other organisations, thus laying the groundwork for new services based on information from multiple sources. In Italy, the Public Administration National Network project was recently launched to specify such a network. However there are currently no enquiry systems available that access huge distributed databases, where data needs to be integrated, interpreted, accessed and presented in natural language.
TAMIC-P will address this problem by supplying a powerful interface that the clerk in a pension organisation can use without knowing the structure and protocols of the many different databases involved. The system will combine natural language technology with menus and graphics to enable coherent dialogues that include searching for a specific item of data. Since technical skills in using a computer will not be necessary, the most important quality of a clerk will become his ability to relate to citizens and their problems. The main goal is to provide a natural language interface to large distributed databases, with transparent access to the underlying data structures, so that customers can have their questions about pensions and social security answered efficiently and effectively.
The two main research fields which have a strong bearing on data access in TAMIC-P are database integration, and the integration of knowledge representation systems (KRS) and databases (DBMS).
The consortium was formed to ensure strong user involvement and tight interaction between users and suppliers. The two public institutions will be the major project users. INPS has 200 agencies nation-wide in Italy with a vast amount of data is stored at several levels: local data in departmental computers, data of national interest in regional or central mainframes. SVB has branch offices in nearly all of Austria's provincial capitals. INPS is the largest Italian public entity in its field serving 19 million working citizens and providing over 14 million pensions, while SVB provides nearly 200,000 pensions. A third user has been set up to co-ordinate public administrations in Italy and will promote the results of TAMIC-P serving as a focus point for the user group.
The system, by reducing the need to train personnel in different software and systems, reduces costs and allows easier operator interchangeability. Operators can give their full attention to the public, who will be presented with a higher quality and transparent information service giving complete information on all social insurance issues.
The operator's role will be transformed from a simple 'information provider' to 'citizen counsellor'. Furthermore, information points can be spread throughout the country, which is especially useful for inaccessible areas.
System performance will be evaluated throughout the life of the project by clerks in their everyday interactions with citizens.