The Austrian Association for Materials Management, Purchasing, and Logistics (BMÖ) has awarded the Austrian Supply Excellence & Digital Procurement Award for the 22nd time. Outstanding solutions in purchasing and supply chain management across the supply chain were honored. The jury evaluated the degree of innovation (innovator role, broad impact), execution (project management, team building, role of top management, etc.), cost-benefit aspects, and transferability to other companies/organizations.

Submitted Process:
“Automated Processing of Order Confirmations with Generative AI in the AGRANA Group”

Jury Justification:
“The processing of order confirmations (OCs) is a key step in the procurement process to ensure the alignment of orders and deliveries. As OCs often come in different formats, languages, and layouts, this process has traditionally been manual, time-consuming, and prone to errors. By using Generative Pre-trained Transformer, AGRANA was able to automatically extract and process data from unstructured documents, regardless of the supplier. 398 documents were processed with a success rate of 95%. The cost of the AI technology per document is significantly lower than manual processing or traditional OCR technologies. This promising result allows for scaling the solution to integrate additional suppliers, order confirmations, and systems.”

The Implementation:
All order confirmations were centralized and forwarded to a central mailbox. A software robot then processes all incoming emails. In the next step, it hands over the order confirmation documents to Generative AI to extract all relevant data. The bot then compares the extracted data with the corresponding order data from the ERP system. If deviations exceed certain thresholds, user interaction is required. Otherwise, the OC process is fully automated.

We consciously chose the use of Generative AI instead of classical AI because it enables the processing of all types of order confirmation documents with a high level of automation.

In close collaboration with the purchasing team, we ensured that the developed automation meets all necessary requirements and optimally supports daily operations.

The automation has been live for several months and successfully supports AGRANA. Together, we plan to further expand its functionality, add new features, and deploy the solution in other departments of the company.