The task I chose DocuPipe for was that of converting pdf’s containing scanned images into Microsoft Access tables. The pdf files contain the results of a graveyard survey – a list of graves with descriptions of the status of the headstone together with the inscriptions. Previously, I was using Adobe OCR to convert pdf’s to Microsoft Word and then using VBA to open a Word file and extract the inscriptions paragraph by paragraph. After editing the inscriptions to remove all the OCR glitches, I then went through manually compiling an Excel file index. The new system requires just one pass of the pdf to create a JSON format file. I then use VBA to process the JSON file assembling two Access tables of inscriptions and index. Much more straight-forward and more practical to hand over to someone else eventually.

Most document-processing systems can handle clean, computer-generated tickets, but our real challenge is handwritten material. Some of our haul-off tickets are extremely messy and hard to decipher, even for people who work with them regularly. DocuPipe consistently converts these documents into accurate, structured data. It can also interpret the many different ways hauler names are handwritten and still match them correctly to the appropriate hauler. That capability is what truly sets DocuPipe apart from the other systems we’ve evaluated.

I find DocuPipe to be very stable and robust, offering high quality results across a large and varied corpus of documents. It's a fundamental part of our system, handling the "heavy lifting" data extraction part with the help of AI agents specifically for the MRO industry in aviation. The cost-effectiveness is another highlight for me. The team behind DocuPipe is incredibly responsive, providing valuable and fast answers to my questions, and they fix issues quickly. Setting it up was super simple, just taking a couple of hours, thanks to the great documentation.