Bulk Purchase Orders with AI and Machine Learning

Background

A leading company in the healthcare sector handles a high volume of purchase orders, each with more than 500 line items, received in various formats such as images, photos of printed documents, disorganized Excel files, and emails. This scenario created a need for manual processing that took nearly a full day per order, slowing operations and affecting customer response times. It was essential to find a solution that automated this process and ensured proper integration with its ERP system.

Challenges

Accuracy and security: It was essential to ensure that only personnel with the proper permissions and certifications could access operational areas, minimizing risks and complying with safety regulations.

Results

The implementation of an innovative solution that combines computer vision for reading unstructured documents and machine learning for automatically mapping product codes resulted in:

- Reduction in processing time: What used to take a full day is now done in minutes, significantly speeding up the workflow.
- 95% automation of the process: Most purchase orders are processed without human intervention, ensuring efficiency and reducing errors.
- Seamless ERP integration: The product codes in the orders were correctly mapped to the system’s SKUs, ensuring accuracy in inventory and operations.
- Improved response times: The company can now process more orders in less time, improving the customer experience and speeding up product delivery.

This project demonstrated how the strategic use of advanced technologies can transform complex, manual processes into optimized workflows, significantly improving operational efficiency.