Highly automated document pro­ces­sing in receiv­a­bles manage­ment

EOS Deutscher Inkasso-Dienst, a spe­cia­list in receiv­a­bles manage­ment, processes around 2.5 million incoming documents every year – supported by AI solutions from Insiders. The flexible adap­ta­bi­lity and con­ti­nuous impro­ve­ment of the solutions are a clear advantage in the constant change and growth of the debt coll­ec­tion service provider. With the new Inte­gra­tion Layer for OCR (ILFO), EOS can now handle even par­ti­cu­larly complex documents and further advance dark pro­ces­sing.

The EOS Group, a company of the Otto Group, is a leading tech­no­logy-based investor in research port­fo­lios and an expert in the pro­ces­sing of out­stan­ding receiv­a­bles. With more than 6,000 employees in over 20 countries, EOS has been sup­porting companies in a wide range of indus­tries in receiv­a­bles manage­ment for more than 50 years, enabling them to get paid for the goods and services they provide and focus on their core business. In Germany, EOS Deutscher Inkasso-Dienst processes millions of documents of all kinds—from struc­tured invoices to freely for­mu­lated emails to hand­written documents.

DIFFERENT DOCUMENT TYPES AND STRINGENT DATA PRO­TEC­TION REQUI­RE­MENTS

Due to the large number of document types and stringent data pro­tec­tion requi­re­ments, document pro­ces­sing in debt coll­ec­tion poses a par­ti­cular challenge. First, the company is con­fronted with a very hete­ro­ge­neous influx of documents – around 80 percent of which are unknown, unstruc­tured documents.

The first step is therefore to correctly identify the document class. This deter­mines which data can be extracted from the document and enables the process to be routed to the right team with the necessary skills for pro­ces­sing. Errors should not occur during this clas­si­fi­ca­tion in order to avoid addi­tional work and longer pro­ces­sing times.

Nutzen für den Kunden

  • Reduktion der manuellen Nach­be­ar­bei­tung um die Hälfte
  • Deutlich ver­bes­serte Hand­schrif­ten­er­ken­nung und auto­ma­ti­sierte Ver­ar­bei­tung
  • Präzise auto­ma­ti­sierte Klas­si­fi­ka­tion reduziert Lauf­zeiten und Mehr­auf­wand
  • Fle­xi­bi­lität durch Inte­gra­tion neuer OCR-Tech­no­lo­gien und indi­vi­du­elle Anpass­bar­keit

“We really like that Insiders proac­tively approa­ches us with sug­ges­tions for impro­ve­ment and that new tech­no­lo­gies are con­stantly being added to the products. In addition, the solutions are so flexible that we can quickly and easily adapt them to our requi­re­ments.”

Stefan Heuer

Product Owner, EOS Deutscher Inkas­so­dienst GmbH

chal­lenges

  • Viel­fäl­tige Beleg­arten, inkl. hand­schrift­li­cher und schlecht gescannter Dokumente
  • Immer neue Use Cases und Beleg­arten
  • Hohe Daten­schutz­an­for­de­rungen
  • Hoher Aufwand für manuelle Doku­men­ten­ver­ar­bei­tung

“Hand­written notes on graph paper – in the past, it was almost impos­sible for us to achieve accurate reco­gni­tion here. With the new Inte­gra­tion Layer for OCR from Insiders, we are achieving signi­fi­cantly better results.”

Stefan Heuer

Product Owner, EOS Deutscher Inkas­so­dienst GmbH

Special attention is required when iden­ti­fying data pro­tec­tion requests in which a customer requests infor­ma­tion about or deletion of their data in accordance with the EU GDPR. Such requests must be distin­gu­ished from general infor­ma­tion on data pro­tec­tion, for example in email signa­tures. Since data pro­tec­tion requests are subject to response deadlines, a high degree of precision is required here.

In the past, the pro­ces­sing of hand­written documents or hand­written notes on documents, poorly pho­to­gra­phed receipts, or scanned incoming mail with imprecise grayscale tones often had to be done manually—with the cor­re­spon­ding personnel costs and pro­ces­sing times. The promised impro­ve­ments offered by new OCR tech­no­logy were therefore more than welcome.