Fritz-Kola: AI consulting for automated invoice processing in SAP
Huseyin Olmez
COO & Co-Founder at Nodient
Last updated: October 27, 2026
In Brief→
The challenge
Fritz-Kola processes more than 3,500 invoices a year for point-of-sale materials, such as fridges and parasols placed in hospitality venues. These invoices were largely booked manually in SAP. The existing OCR solution only read the header data of an invoice, and an earlier automation proved fragile as soon as suppliers changed their invoice format. The IT team was looking for a structural solution that fits the existing systems.
The solution
Together with the IT team and the field service department, Nodient analyzed the entire invoice process, from incoming email to booking in SAP. Based on this, we worked out several integration scenarios in which AI complements the existing systems rather than replacing them. Our advice describes how invoice lines can be read reliably and processed automatically. It was also assessed against GDPR and information security requirements, with AI processing within the EU.
Our approach
The project started with conversations with the teams involved and a detailed walkthrough of the current process. We then captured the bottlenecks in a decision tree and a gap analysis, compared the architecture options and calculated the costs. The advice was developed into a phased implementation plan, with a proof of concept as the first step.
Outcomes→
3,500+
invoices per year analyzed
5
integration scenarios worked out and compared
175+
hours of manual booking work per year mapped
