Process AP Delivery Receipts with Rapid Adapt
The “Process AP Invoices with Rapid Adapt” workflow automates the ingestion of AP Invoices from a hot folder and extracts key index data from the document. Using GlobalCapture’s Rapid Adapt Forms builder, the first document through the system will end up in validation and require the user to leverage a point and click designer to assign marker’s used to identify the document and the positional location of the data points to be extracted. Going forward, most subsequent documents should get classified and automatically processed. Because OCR is never 100% accurate, it should be expected that misreads will happen and a small percentage of future documents will require validation. The quality of the original, quality of the scan and the font type being read will have the biggest impact to OCR accuracy.
The validation process in this workflow leverages Forms Learning via RapidAdapt. For details on how validation users interact with RapidAdapt, refer to the documentation for the feature here.
For demo and testing purposes, click here to download an archive of example invoices to be used with this workflow.
Understanding the Invoice Transformation Process
To incorporate this workflow into your business process and make edits or perform maintenance, it is important to completely understand how each node comes together. As such, a node by node review of the as-built workflow is outlined below. Users that have taken the Certified GlobalCapture Engineer (CGCE) eLearning and Advanced GlobalCapture Engineer (AGCE) education class will have the foundation of knowledge necessary to understand and implement this workflow.
The green Import Node of the workflow determines the location GlobalCapture will monitor to automatically ingest the documents and the frequency at which it does so.
- Change the “Source Path” attribute to change the folder GlobalCapture monitors and imports from, along with the “Frequency” attribute to determine how often GlobalCapture performs in import. More information on all Import node attributes can be found in the GlobalCapture Workflow Management documentation.
The first blue Set Process Field Node sets workflow default parameters. The first is a flag used to determine if any page separation will be enabled. Page separation is used to split pages when multiple individual documents are contained in a single file placed in the hot folder.
- To split on a static page count, set the Separation Flag to 2. This will configure GlobalCapture to separate every X pages as configured on the first blue Separate Node.
- To split using a QR Code or Datamatrix barcode coversheet that appears between each individual document in the file, set the Separation Flag to 1
- To skip separation and assume each document in the pick up location represents a single invoice, set the Separation Flag to its default value of 0.
The second parameter sets the AP Record Type field to Vendor Invoice.
The next set of nodes contained within the Separation Configuration group control how documents being processed will be separated, according to the separation flag set earlier. To avoid issues, changes to these nodes should only be made by advanced users with a deep understanding of how separation works.
The first purple Condition node checks if page count separation will be used as configured in the Separation Flag previously.
If the flag is set to two, the workflow will proceed to use page count separation.
The first blue Separate node is used to configure page count separation. The Page Count attribute determines the number of resulting pages in each document. A value of 2 will initiate separation after every page in the document. If not the workflow moves to the next purple Condition node.
The second purple Condition node checks if barcode separation will be used as configured in the Separation Flag previously. If the flag is set to 1, the workflow will proceed to use barcode separation.
The second blue Separate node is used to configure barcode page separation. GlobalCapture supports many barcode formats for use in separation. The default configuration uses QR code and DatMatrix formats. Additionally, so as not to initiate a false positive, a prefix can be used that reads the value in the barcode to find a unique value only intended for separation. The default prefix looks for the value “SEPARATOR” and will only initiate separation when it is found.
- For more on separation, consult the Separate Node documentation.
- The brown Delete Pages node is used to remove pages from the resulting document(s). When using barcode cover sheets for separation, once the sheets have served their purpose the sheets can be deleted from the resulting documents, saving space. The workflow is configured out of the box to delete barcode pages that include a QR code with the value "SEPARATOR".
The third purple Condition node checks to ensure there are still pages remaining in the document using the Page Count system field. This is best practice whenever using the delete pages node as downstream errors can occur when trying to process a zero page document.
If pages remain the workflow moves straight to the classify node.
- If no pages remain, the red End node is reached and the workflow ends. A common example of where this can occur is when two barcode coversheets are mistakenly placed on top of each other. A resulting document would contain only the barcode cover sheet which would get deleted, leaving no pages.
The Classify node is used to initiate OCR extraction on the document with predefined templates.
- If the document matches an existing template or template group that has been applied, it will use that templates rules to extract data and move on to the release node.
- The first time a document is processed through this workflow, no templates will exist and the document will move to a validation queue in Batch Manager where a user can build an extraction template that matches that documents format.
When leaving the Classify node, documents that match an existing template will follow the Classified path.
A number of data clean up rules will work on the Amount field in the next blue Set Process Field Node. These rules are in place to normalize data extracted via OCR and include removing spaces, dollar signs, decimals read as commas, and commas read as decimals.
The orange Release Node is used to release the document into your GlobalSearch Business Essentials database Accounts Payable archive.
The red End node signifies the end of the GlobalCapture process.