Achieving 90% Accuracy in Document Processing with AI for the Port of Rotterdam Authority

V G • Dec 02,2024

Summary:

By achieving 90% accuracy with just 10 training documents, the Port optimized its operations, enhanced compliance, and set a benchmark for leveraging AI in logistics and procurement management. 

Client: 

The Port of Rotterdam is one of the largest and most technologically advanced ports in the world. 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. 

Problem Statement: 

PoR collaborates with a wide range of suppliers, each submitting invoices with diverse layouts. The Accounts Payable (AP) team handles approximately 20,000 invoices annually. While they previously relied on legacy OCR software, the system required technical teams to create templates for each supplier, slowing down the onboarding process. Additionally, the AP team faced inefficiencies due to manual workflows and challenges caused by the significant number of non-purchase order invoices, which demanded precise line-item extraction to avoid errors.

Managing the complexity of logistics, documentation, and compliance while ensuring operational efficiency requires sophisticated systems. 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.

 

Results: 

  • Achieved 90% accuracy in document processing after being trained on just 10 documents.
  • A 70.7% reduction in time needed to process invoices.
  • Saved substantial costs associated with manual labor, error correction, and compliance breaches.

AI Solution Overview:

To address challenges in document processing, the Port partnered with Rossum AI, an intelligent document processing (IDP) platform, to automate and streamline its operations.

PoR utilized Rossum’s advanced proprietary LLM, which has been trained on millions of transactional documents to enhance its contextual comprehension. 

The Rossum team worked with PoR to customize the AI outputs to their specific business requirements and workflows. Rossum AI’s pre-trained models required only 10 documents to adapt to the Port’s unique documentation needs.

PoR implemented Rossum’s Signature Onboarding hypercare plan, specifically designed for large, complex enterprises. Business users received training to navigate the interface, enabling seamless human-AI collaboration even for intricate data extraction tasks. 

Once users gained confidence in the platform’s capabilities, PoR adjusted the system’s confidence thresholds to enable touchless, straight-through processing.

To establish a comprehensive workflow, PoR collaborated with Rossum’s Professional Services team. They ensured a secure and reliable integration between Rossum and the SAP ECC system, as well as the Source-to-Pay procurement solution. The team deployed a Rossum extension to pull master data from SAP, which was then used to validate extracted information, ensuring data accuracy in the downstream system of record.

References: 

  1. How the Port of Rotterdam Authority achieved 90% accuracy after only 10 documents. https://rossum.ai/customer-stories/port-of-rotterdam/?_gl=1*19265qz*_up*MQ..*_gs*MQ..&gclid=CjwKCAiA0rW6BhAcEiwAQH28ItyWPaSmdXNVWLnTP8v8-iF_oAkE855FQmSdJhbGfpvTwb2kjs8s2hoC330QAvD_BwE
  2. 10 AI Procurement Use Cases & Case Studies. https://research.aimultiple.com/ai-procurement/#easy-footnote-bottom-2-179442

Industry: Transport Services

Vendor: Rossum

Client: The Port of Rotterdam Authority

Previos Article

Next Article