Oldcastle APG, one of the leading manufacturers in the global architectural products industry, has modernized its proof of delivery (POD) document management process following a collaboration with Amazon Web Services (AWS). Previously, the company used an optical character recognition (OCR) system that was ineffective and required significant maintenance, achieving only a 30 to 40% accuracy in reading between 100,000 and 300,000 documents monthly.
The Oldcastle team faced various challenges, including the need to process a large volume of documents with minimal human intervention, dealing with diverse formats, and providing instant visibility into pending PODs. In over 200 facilities, dispatchers spent between 4 and 5 hours daily manually processing these documents, causing delays and increasing the potential for errors. To address these issues, Oldcastle collaborated with AWS solution architects to develop an innovative workflow.
The new document processing architecture is based on Amazon Textract and Amazon Bedrock. This allows the company to send shipping tickets directly through Amazon’s email service. Documents are now automatically processed through a series of steps that include document analysis and key data extraction. This upgrade has not only automated document handling but also improved signature validation and data extraction accuracy.
As a result of this implementation, Oldcastle has achieved significant advancements in its business process. The need for manual intervention has been reduced, data extraction reliability has increased, and customer satisfaction has improved. The company has enhanced its document processing capacity, moving from a few thousand to a range of 200,000 to 300,000 documents per month, without experiencing performance issues.
Additionally, Oldcastle is considering expanding this solution to other uses, such as processing vendor invoices and validating W9 forms. This strategy not only optimizes current operations but also positions the company for future growth and greater efficiency.
Source: MiMub in Spanish
