Streamlined document processing and 15,000 man-hours saved per year.

• ChallengeAutomate document processing with OCR
• SolutionOCR to cut time on invoice processing
• Technologies and toolsPython, Pytorch, CatBoost, scikit-learn, OpenCV, Tesseract, Google Vision API

Client

AsstrA is an international transport and logistics company. They were facing the problem of image recognition and manual data processing. Manual document processing and data extraction have been taking up the bulk of their time. More importantly, it was rather costly and often led to undetected errors. That’s why they wanted to remove the human element from document-based processes.

To accomplish this, the client approached InData Labs, OCR solutions provider.

Challenge: automate document processing with OCR

The client was bogged down with manual document preparations. That’s why they asked us for an efficient OCR solution to improve the invoice data capture procedure in the workplace.

The biggest challenge for our team was to develop a system that is able to extract data from thousands of invoice templates and requires minimum human involvement to adapt to new templates.

Solution: OCR to cut time on invoice processing

Logistics is ripe for OCR application. OCR technology has been proven effective in identifying such shipping documents as airway bills, bills of lading, invoices, CMRs, import licenses and so on.

Our team started with classifying various document scans including invoices and CMRs. Once the document scan was classified, we had to check, if there were any errors or anomalies in the document scans. Our team also had to find and identify seals on documents. To do this, we used CV (Computer Vision) technology.

Another task was to recognize and derive the necessary data from like invoice date and invoice number.

The approach we used during the project:

Accountants used to spend much time on document processing. Our OCR solution aims to upload and process documents on its own. What’s more, it registers recognized documents automatically. The solution processes one document four times faster than a human does. According to our analysis, automated processing and OCR invoice capture of 300,000 documents will save 15,000 man-hours per year.

As a result, our OCR invoice software helps the client run the business more effectively, reliably and predictively.

Result: improved invoice processing and reduction in man-hours

Before partnering with the InData Labs team, the client had to process the wealth of data manually on a regular basis. Having plenty of experience in developing OCR invoice software, our team has aided the logistics company with an intelligent customized solution that streamlines document processing.

The client has benefited from cooperation with us in the following ways:

• automated document processing and registration frees up the time
• the risk of human error in the workplace lowered
• increased productivity
• reduced man-hours
• reduced labor costs