P2P-O: A Purchase-To-Pay Ontology for Enabling Semantic Invoices

Michael Schulze; Markus Schröder; Christian Jilek; Torsten Albers; Heiko Maus; Andreas Dengel

In: Ruben Verborgh; Katja Hose; Heiko Paulheim; Pierre-Antoine Champin; Maria Maleshkova; Oscar Corcho; Petar Ristoski; Mehwish Alam (Hrsg.). The Semantic Web - 18th International Conference. Extended Semantic Web Conference (ESWC-2021), June 6-10, Hersonissos/Virtual, Greece, Pages 647-663, Lecture Notes in Computer Science, Vol. 12731, Springer, 2021.


Small and medium-sized enterprises increasingly adopt electronic invoices and digitized purchase-to-pay processes. A purchase-to-pay process begins with making a purchase order and ends with completing the payment process. Even when organizations adopt electronic invoices, knowledge work in such processes is characterized by assimilating information distributed over heterogeneous sources among different stages in the process. By integrating such information and enabling a shared understanding of stakeholders in such processes, ontologies and knowledge graphs can serve as an appropriate infrastructure for enabling knowledge services. However, no suitable ontology is available for current electronic invoices and digitized purchase-to-pay processes. Therefore, this paper presents P2P-O, a dedicated purchase-to-pay ontology developed in cooperation with industry domain experts. P2P-O enables organizations to create semantic invoices, which are invoices following linked data principles. The European Standard EN 16931-1:2017 for electronic invoices was the main non-ontological resource for developing P2P-O. The evaluation approach is threefold: (1) to follow ontology engineering best practices, we applied OOPS! (OntOlogy Pitfall Scanner!) and OntoDebug; (2) to evaluate competency questions, we constructed a purchase-to-pay knowledge graph with RML technologies and executed corresponding SPARQL queries; (3) to illustrate a P2P-O-based knowledge service and use case, we implemented an invoicing dashboard within a corporate memory system and thus enabled an entity-centric view on invoice data. Organizations can immediately start experimenting with P2P-O by generating semantic invoices with provided RML mappings.


Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence