Pupuk Indonesia successfully streamlines Data Extraction Process time 

The Challenge

PT. Pupuk Indonesia have numerous documents that contain a variety of unrelated design elements such as watermarks, irrelevant images, and other distracting features. Laying out and organising these contents can be a significant challenge, as there is currently no optical character recognition (OCR) technology available. The lack of an effective OCR solution makes it difficult to quickly and neatly structure the content of these documents in a clear and coherent manner. 

Moreover, this is an obstacle for organisations and individuals who need to process the information contained within these visually cluttered documents in a streamlined and productive way. Developing more advanced OCR capabilities that can handle documents with diverse and unrelated design elements would be highly beneficial.

The Solution

To solve the problem of Scannable PDF data extraction, Devoteam uses Vision AI and Gemini to get the information needed by Pupuk Indonesia related to employee profile data. Vision AI reads each document, which will later be transformed by the system built by Devoteam to adjust their layouts. Meanwhile, Gemini acts as an NER to get employee profile data. In the implementation stage, challenges related to data with too many decorations and watermarks caused a decrease in Vision AI performance.

In addition to the document cleanup functionality, Devoteam also created a comprehensive layout management system. This system utilises X and Y coordinate data in document images to automatically organise content in a clear and coherent manner. By basing layouts on the actual spatial information of the document, rather than relying on predictive machine learning models, the resulting layouts are highly accurate and do not require extensive training or fine-tuning. This simplified approach ensures that documents can be organised quickly and efficiently, making the information they contain easy to access and use.

The collaborative efforts between Devoteam and Pupuk Indonesia, leveraging the power of Google Cloud Platform, have resulted in a powerful solution to overcome the challenges associated with processing complex documents that are burdened by unwanted design elements and layout issues. This innovative system not only improves speed, but also accuracy and efficiency in document validation that is impacted in increasing productivity.

The Result

The Pupuk Indonesia team has successfully implemented the custom document cleaning and layout management system developed by Devoteam, leading to significant improvements in productivity and cost savings.

It would typically take the Pupuk Indonesia team 5 to 10 minutes to collect the necessary information from a single document. However, with the new automated system in place, this data extraction process has been reduced to a mere 40 to 70 seconds per document. The Google Cloud Platform infrastructure provided the scalability and processing power needed to handle the high volumes of documents efficiently.

Furthermore, the new system has also led to substantial cost savings for Pupuk Indonesia. With the automated system handling the majority of the work, only a single person is needed to validate the results and address any unfamiliarities that the system may flag. This reduction in human resource requirements has resulted in significant cost savings for the organisation, allowing them to reallocate those resources to other critical business initiatives.

Thus, the successful implementation of Devoteam custom document cleaning and layout management solution, powered by Google Cloud Platform, has been a transformative experience for Pupuk Indonesia. By streamlining their document processing workflows, the organisation has been able to enhance productivity, reduce operational costs, and ultimately improve the overall efficiency of their data-driven decision-making processes. This strategic partnership with Devoteam and the utilisation of Google Cloud Platform have positioned Pupuk Indonesia as a leader in leveraging innovative technology to optimise their operations and gain a competitive edge in the industry.