AI-based Document Analysis

By implementing an AI solution, SPIE can automate the maintenance order process, eliminating manual naming, sorting, and time tracking. This significantly reduces processing time and lowers costs.

Previously, the maintenance order handling at SPIE was a manual and time-intensive process. Digitally submitted maintenance orders had to be processed manually, which was prone to errors and demanded considerable resources. To overcome these challenges, the project was initiated to develop a more efficient and contemporary solution.

Within this project, an AI-driven system is deployed to automate workflows starting from the receipt of completed maintenance reports. The AI analyzes and categorizes scanned maintenance documents that were formerly handled manually. It autonomously differentiates between customer and SPIE documents and extracts critical information such as order numbers, execution periods, and logged working hours. The system subsequently generates export files to facilitate downstream processing. Documents that cannot be conclusively interpreted are directed to an error queue, which also serves as feedback to continuously refine and train the AI model.

The core innovation of this use case lies in the end-to-end automation of a previously manual procedure through advanced artificial intelligence, resulting in substantial time savings, increased accuracy, and the elimination of reliance on external service providers.

Currently, the solution is undergoing testing and has demonstrated promising outcomes. The automation has accelerated process KPIs and alleviated the workload of staff in the order center. The newly implemented AI solution optimizes resource allocation while enhancing the overall quality of data processing. This milestone represents a significant advancement towards a digitized, efficient, and future-ready operational framework at SPIE.

Duration:
Test run May 2024 - ongoing

Contact:
Stephan Streckwaldt
stephan.streckwaldt@spie.de

Artikel teilen