• mailRetrieval is an expert system for implementing error-tolerant search and automatic duplicate recognition and clustering in large address databases. With this system it is possible to find addresses or to recognize them as duplicates, even when the data which has just been entered differs from the data stored in the database because of reading, hearing or data entry errors or because of different abbreviations (use of synonyms or differing recording rules). This takes place via the analysis of name and address information with the aid of country-specific rule and term tables, language area specific phonetics and address-specific fuzzy logic.
• mailRetrieval is designed for 1:n comparison when the objective is to implement an error-tolerant search in real time for an address in a very large address database. This procedure can be done smoothly – even for a corresponding large number of users – several hundred times per minute with different addresses for each user.
  Even with today's computer capabilities, this problem can only be solved in an acceptable length of time in a high and reliable quality with software technology that is especially designed for it.
• mailRetrieval enables many simultaneously active users to operate even the largest databases with the highest recognition quality and with reponse times of less than one second. Of course, the particular performance is dependent on the hardware that is used and on the particular task at hand. But even with rather small INTEL servers, for address databases in the range of 5 – 10 million addresses, average response times of < 0.5 seconds are attained.
• In addition to real time name and address search and the function duplicate recognition, mailRetrieval offers a real time update function. An address that has just been entered is available in the next instant to all users for an error-tolerant search. Of course data and transaction integrity are observed. In this age of multi-channel marketing, this is the only way to generate a Single View of Customer.

• mailRetrieval does not run (offline), but rather is directly integrated into your applications and processes. Users also do not have to "cut" and "paste". You keep your usual screen masks and merely receive several additional messages or selection lists. Hence, mailRetrieval becomes an integrated function of your application in the background that enables users to search quickly and in an error-tolerant way for stored addresses and to ensure automatically in real time that duplicates never even arise.
• Hence, not only can all name and address elements be compared with each other in an error-tolerant way, but also fields closely related to the address such as telephone number, e-mail address, WEB address, date of birth, etc. In addition there are also a number of empty fields in which you can also include other information such as contract, account or credit card number etc. in the comparison. All comparison fields can be evaluated with special - and for the particular information type specific - error-tolerant comparisons.
• The system is highly flexible. You yourself can adapt the system exactly to your requirements via extensive parametrization options. You determine on your own under which conditions an address which is to be entered or changed should be recognized as a duplicate and in which case this is either a certain or an uncertain duplicate. Or which clusters will be dynamically generated and how this information will be placed in the database and used in additional applications.

• mailRetrieval is easy to integrate into your application completely independent of whatever system environment or with which program language your development takes place in or uses. A multitude of well-designed and proven interfaces are available for this purpose. You therefore need not invest in the development of "bridge programs" or "connectivity software" in order to use mailRetrieval. For mySAP.com a special "ready-to-run" solution is available, which in many cases requires absolutely no additional development effort and expenditure (Solutions for mySAP™ Business Suite.
• Special country and platform versions are available for consumer addresses (B2C) and for business addresses (B2B).
• Upon rollout the system already has several standard parameter sets for applications. For that reason, important typical requirements for the application and for the respective country are already covered. Extensive "trial and error and experimentation" is not necessary.

• The expert system mailRetrieval is available for all generally used programs ranging from Windows to UNIX and iSeries (AS/400) to Mainframe.
• It does not presume a special database system, but rather operates together with the generally used database systems and transaction monitors of the particular platform.
• mailRetrieval is completed in client/server architecture and can be arbitrarily distributed in your system environment. For example, the mailRetrieval server can run on the same computer as your application or your database or on a stand-alone computer.
• Because of the error-tolerant search method that was especially developed for address management in real time, you dont need a "super-computer" to ensure acceptable performance for very large databases.
• Because of the availability of a multitude of country-specific versions, mailRetrieval can also be smoothly integrated into international applications. The consideration of country-specific particularities is self-evident.