Robotic process automation technologies or RPA, are mostly used to replace manual tasks with others that can be executed automatically.
This includes activities such as file management, scan, open and sends emails, get and enter information from/into applications, get data from documents, manage spreadsheets, etc.
RPA FOR APPS INTEGRATION?
A nontraditional use of RPA technology is applications integration, where this type of tools allows applications to be integrated by replacing the manual steps that would be required to get the data from one application and entering it into another application.
When you develop a programmatic integration to link to applications, complex interfaces, and integration processes have to be built using different types of technologies. And there are many choices from simple text file sharing to complex SOA based architectures. All of them require a significant development effort that provides big benefits in terms of processing speed, reliability, etc.
When the volume of information to integrate is considerable, the traditional integration approach es the most appropriate. But when integrated data is not updated frequently, the use of RPA can be a good choice to easily integrate applications.
RPA technologies allow processes to be defined that mimic human behavior while operating applications. You can build a process that logs into an application, navigates to a specific window, query an invoice –for example-, completes the information with data collected from other sources (like a spreadsheet for example, if invoices distributions are calculated from another source), then connect to the target application and enter the complete invoice with the information collected from different sources.
In a later step, we could even run a process in the target system to generate the invoice print in pdf format, and then send an email to a customer with the generated file. All of this can be built with an RPA process that simulates the manual steps to complete these activities.
IS THIS AN EFFICIENT PROCESS?
For high volume integrations, this is definitely not efficient. But for integrations that require a small volume of data, this is a really an interesting alternative especially for organizations where RPA technology is already established.