How a DMS Company Implemented Cognitive Invoice Data Capture
Our client, a DMS provider, specializes in electronic data interchange (EDI), which is a system for paperless exchange and processing of structured documents such as orders, invoices (and many others) in electronic form. In addition, our client runs its cloud-based Document Management System for a large group of clients.
Our client implemented Rossum to process all PDF and scanned invoices, which were previously processed manually. This saves time without having to implement any templates, and improves accuracy thanks to the AI.
What was happening?
Our client ran a system for processing EDI invoices and other documents for a large group of clients. All of the invoices came in widely varying formats and it was a huge task to process them. Their clients could process EDI invoices with little human interaction. For smaller businesses, which work mainly with PDF and scanned invoices, they had to be handled manually. Template-based OCR systems could not be used for these documents due to the wide variability of invoice layouts. Their solution thus failed to cover a significant portion of the use cases for invoices, particularly from the SME segment. To offer a full-scope automation service, they needed a different way to extract the information from EDI invoices and PDF/scanned invoices – this is where we stepped in.
How Rossum worked with the DMS provider
Our client implemented Rossum, an invoice data capture tool, which can work without any templates thanks to Rossum’s artificial intelligence technology. Now they can process all types of invoices without the need for manual data entry.
The DMS provider’s clients continue to automatically send PDF and scanned invoices to Orion via email. Orion now sends pictures and PDFs into Rossum’s Data Extraction REST API, where Rossum extracts the selected fields from the invoices. There is no setup of templates or rules needed to extract the information – the app just runs the documents through its data capture and OCR neural networks and collects all captured fields in a JSON structure. Next Rossum sends back the JSON file with the extraction results, and Orion displays the extracted data in the individual client accounts. Finally, the data is sent into individual clients’ ERP systems.
How the DMS provider benefited from Rossum
By implementing our API into their DMS, our client was able to automatically extract data from scanned and PDF invoices. The following benefits were immediate:
– No templates: Rossum does not require a lengthy and expensive set-up of rules or templates.
– Ability to extract invoices from SME: Rossum can extract data from invoices with very different layouts and formats with human-level accuracy, even if the invoice is in a previously unseen format.
– Smooth integration: We offer a range of APIs that are easily integrated into any DMS system.
By implementing Rossum, the DMS provider can now offer complex services for invoice automation on top of their EDI competencies. Without the need for manual data entry of invoice data fields, it not only saves time but also money, adding benefits in all aspects of their business. Invoice processing can now be done for both large corporations and SMEs, as there is no need to set up templates thanks to our AI.