How the Port of Rotterdam Authority achieved 90% accuracy after only 10 documents

Port of Rotterdam case study

Who are they?

The port of Rotterdam is the largest maritime cargo port in Europe. The aim of the Port of Rotterdam Authority (PoR) is to strengthen the competitive position of the Port of Rotterdam as a logistics hub and a world-class industrial complex in terms of both size and quality. PoR is focused on accelerating sustainability in the port and is a partner in the digitalization of the port and logistics chains. The increasing complexity of PoR activities, involving multiple parties and a high pace of throughput, requires an entrepreneurial and effective organization.

What was the challenge?

PoR works with many different suppliers each submitting invoices with a wide variability of layouts. On average PoR’s Accounts Payable (AP) team receives 20,000 invoices per year. Although they already had legacy OCR software in place this was reliant on technical teams to set up templates, which impacted supplier onboarding. The efficiency of the AP team was further reduced as its processes were largely manual and a high percentage of non-purchase order invoices were causing issues requiring high line item extraction accuracy.

PoR wanted to replace its legacy OCR technology with an AI-based intelligent document processing (IDP) solution. The project goal was to automate invoice processing to increase efficiency, speed up throughput times, reduce human error, and alleviate the pressure on the AP team.

PoR chose Rossum, a transactional document processing platform, for its market-leading AI, integrations and out-of-the-box extensions environment.

“We were looking for a SaaS technology that enabled highly automated invoice processing using intelligent document processing. We chose Rossum for its advanced AI, data quality accuracy, customizability and out-of-the-box integrations for end-to-end automation.”

Arthur Philippa, Lead RPA Developer Port of Rotterdam.

What did Rossum do? AI you can control

The migration to the Rossum IDP solution marked the initial phase of a broader digitalization project. Adoption of Rossum by AP personnel was therefore critical to success, as was the configuration of Rossum to fit PoR’s business rules and processes and integration with downstream systems.

The key objective was to automate the transactional document process end to end. To do this Rossum needed to gain the trust of the business-user team by demonstrating the accuracy of our AI extraction models. It was important for the team to feel in control of the AI and could see it instantly learn at scale from a few inputs.

PoR leveraged Rossum’s proprietary LLM, trained on millions of transactional documents for increased contextual understanding. PoR also adopted a Signature Onboarding hypercare plan, which is recommended for large complex enterprises. The Rossum team worked with PoR to customize the AI outputs to their specific business requirements and workflows. While the business-users were trained to use the interface which enables fast human-AI collaboration, even on complex data extraction tasks. Once the platform had gained the business-users’ trust PoR was able to adjust confidence levels to power touchless straight-through processing.

To deliver an end-to-end workflow, PoR engaged Rossum’s Professional Services team to facilitate a reliable and secure connection between Rossum and SAP ECC, as well as their Source to Pay procurement solution. The team deployed a Rossum extension to retrieve master data from SAP and used it to validate the extracted data and ensure accuracy with the downstream system of record.

Thanks to this and our advanced AI, together PoR and Rossum successfully delivered on three key project objectives:

  • Significant reduction in time spent on previously manual processes.
  • Integration and master data matching with the downstream ERP (SAP ECC).
  • Integration with PoR’s Source to Pay procurement solution (Proactis).

Rossum has reduced the time needed to process invoices by 810 personnel days across the year, this is a 70.7% reduction in manual effort.

Next-generation AI – instant learning a new document use case

Rossum Aurora intelligent document processing logistics OCR

The successful digitalization of invoice processing has encouraged PoR to explore new transactional document use cases to scale automation across the business. For instance, PoR is deploying our next-generation AI in conjunction with its RPA tool for a logistics use case aimed at automating the validation and processing of Tonnage Certificates.

Rossum’s next-generation AI, Rossum Aurora, is powered by our LLM to quickly gain an understanding of each new document it “sees”. Rossum Aurora applied this instant learning to quickly achieve market-leading accuracy levels, it took only 10 Tonnage Certificates for the AI to reach >90% accuracy.

This market-leading accuracy performance has helped build trust in Rossum’s AI, which along with the easy-to-use interface, has contributed to the adoption of Rossum by PoR’s team.

“We are enthusiastic about Rossum’s performance and its capability to quickly learn and adapt to the wide range of Tonnage Certificates used by shipping agencies! This has been instrumental in building the team’s trust and confidence in this AI automation initiative.”

Arthur Philippa.

What’s next?

We are continuing to support the Port of Rotterdam Authority extract the maximum value from Rossum and to scale transactional document automation within their operations.

“We are enthusiastic about Rossum’s performance and its capability to quickly learn and adapt to the wide range of Tonnage Certificates used by shipping agencies! This has been instrumental in building the team’s trust and confidence in this AI automation initiative.”

Arthur Philippa, Lead RPA Developer Port of Rotterdam.

Saving time, improving efficiency


AP days saved per year


Reduction in AP manual effort


Accuracy after 10 tonnage certificates

Ready to get started?

See why hundreds of companies power their document processing with Rossum.