lobo solutions and interfaces

Automatic invoice recognition


The invoice recognition is an advanced solution for the automatic processing of all national and international invoices and credit advices. The invoice recognition interprets, independent of the language, the complete text of the invoice and extracts, with a high level of reading quality, all data relevant to bookings from invoices and credit advices, that would normally be needed in the invoice recording process. The invoice recognition thus perfectly closes the media break in the accounting department.

Benefit aspects
The generic basic set of rules implements, in a robust manner, usually independent of the language, the extraction of the data relevant to accounting from the total of all document types of document that can occur, such that on a total average, optimal results are achieved. The solution is based on many years of experience in various projects and is flexible and easy to use. The efficient configuration of individual special features is possible in many areas.

Via the training mode the invoice recognition is “capable of learning”. If necessary, information not completely recognised can initially be learnt in the invoice typical VERIFY-dialogue. During the data extraction of invoices and credit advices the solution mainly does without the keyword search. As a result, adaptations and extensions to other languages and countries are possible in a short time with little effort.

Invoices with several pages can be processed without any problems, i.e. especially invoices with tables over several pages. Using special read algorithms for the recognition of the tables the solution processes single line table positions and, in many cases also such where the positions show a variable number of lines.

For the safe definition of the creditors the master data stored in the user’s ERP system is used. In ambiguous situations a special technology offer possibilities for differentiation using safe textual characteristics in the document. Further inspections at a position level based on the order data are possible.

Not only paper invoices can be processed using the solution but also electronic invoices. This is possible by the ScanClient or the import components being use implementing an appropriate interface.

The system is optimised for throughput and for reasons of recording efficiency runs batch-oriented. By batch-splitting in connection with a sophisticated resubmission menu it is possible to put individual (problematic) documents on hold for a certain time, without delaying the flow of the remaining documents. This can then be necessary if something needs to be clarified, or if, e.g. at the moment incoming goods information is being waited for. There, where the standards implemented and the extensive configuration possibilities do not suffice, a powerful scripting interface on a data and component level offers unlimited possibilities for individualisation. This is used, among others, by many service providers and VARs.

Functions – invoice handling process


  • Batch-oriented handling
    Normally the documents are processed on pre-sorted batches. The pre-sorting is done using a separator sheet or by a barcode label on the first page of the document. Attachments, which are marked by a separator sheet or a special barcode on the first page of the attachment can be excluded from the extraction. The processing of not pre-separated documents is also possible.
  • Standardised data interfaces
    The standardised data interface supports the import of images, metadata and control information in “flat” and/or “hierarchical” structures. Besides binary images colour images can also be imported. The binary (TIFF-) images are used for reading writing. The colour (JPEG-) images are only passed on through. The handshaking with the passing-on process is done using signal files.
  • Transaction secure data handling
    As a result of the import, all data is buffered in a transaction secure SQL data-base for further processing. The data in the data-base is only then deleted, if the downstream application gives the appropriate signal for deletion.
  • Scripting interface
    The import is capable of scripting and can, in addition to the standard, be adapted to individual circumstances.

Recognition and extraction

  • Flexible supplier determination
    By searching for known account information, VAT No. and the suppliers IBAN as well as using the extendable master data (cross-reference to address information etc.) and the use of the proven XFingerPrint technology.
  • Highest extraction performance
    By using modern character recognition software, also from third parties, in context with the free writing recognition XConText and the leading table finder.
  • Higher blind-bookings rate
    By complete table comparison (comparison of complex table structures, processing of fields with up to 5 decimal points etc.) Controlled training Targeted increase of the reading performance for problematic suppliers; controlled via the rework. Can also be used on table columns (such as article text), hereby the different positions in the first page and on the following pages are taken into account.

This solution is available for





Become curious?

You would like to learn more about our solutions and products?