Accounts Payable Automation in the Cloud

One of the important features of SmartSoft Invoices it that it can easily be deployed in the cloud.

For this deployment we used Microsoft Azure. But AWS, Google Cloud, Oracle cloud, and any of the popular cloud platforms would work too.

The classification station can automatically classify documents based on various criteria. In this case the criteria for document classification is “language”. Documents can be from different countries (US, UK, DE) with their different languages. The document classification engine uses a neural network which basically means that no manual rule setting should be applied, the machine learning approach lets the system learn from provided samples. The more you teach it the better it will perform.

Users can also map Azure storage folders to local drives so documents placed in the corresponding local folder are automatically sent to the system. The batch of documents is then processed by the Data Capture engine to extract key data into the fields of the form. Invoices from separate countries are treated as separate document types where for UK and DE invoices we extract IBAN and VAT ID, which is not relevant for US invoices.

The system is scalable which means that it can do automatic invoice processing of millions of pages per day. A new OCR station can be easily added when processed volume is increased, to help improve the speed of processing. OCR processing is the most CPU intensive part of the process. To minimize costs the system provides the option to automatically turn it off. The user can set the system to turn itself off which comes very handy at nights and weekends.

Locally, on the user’s computer, we need to only install a Verify station for a verify operator to make sure the captured data is correct.

The general advantages of cloud deployment are valid – no need to invest in hardware and no need for local IT personnel to administer the system and the servers. The cloud service provider, in this case being Microsoft, takes care of the hardware and the connectivity.