Hello everyone,
I am working on a crucial automation project involving UiPath Document Understanding and need expert advice on its feasibility for my specific scenario.
My Use Case:
I need to process various Shipping Documents, specifically Air Waybills (AWB) and Bills of Lading (BoL), which come from multiple carriers/freight forwarders. The primary challenge is the high variability in the document layouts (as seen in the provided samples).
I need to reliably extract and tabulate the following critical data points into a structured database:
- AWB/BoL Number (e.g., 2025JLCB0004001, KLID00430626, 080400685603, NTI-NAE09826, JKT5001546)
- Invoice Numbers (Multiple line items, e.g., K/000098/2025, K/000217/2025, D/00077/2025 series)
- Departure/Issue Date (e.g., JAN/22/2025, 16-Jan-25, JAN.20.2025, 14-Jan-2025, 13 Jan 2025)
- Shipper and Consignee Information
My Core Questions:
- Is Document Understanding (DU) with its Intelligent Form Extractor and Machine Learning Extractor the recommended approach for handling these highly unstructured/semi-structured logistics documents with such significant layout variation?
- What is the best practice for configuring the model/extractors to handle the multiple, small Invoice Numbers that often appear as line items or attachments?
- Are there any pre-trained Document Understanding ML Models available specifically for common logistics documents (AWB, BoL) that I should leverage before training a custom model?
Any guidance on the correct sequence of DU activities (e.g., which Classifier/Extractor combination works best for this industry) would be highly valuable!
Please find the sample use case in the attachment
Thank you in advance for your help.
Example :
