Robotic process automation started as a more efficient way to write macros on desktop computers. However, many enterprises had difficulties scaling it beyond a few software robots called "bots."
As interest grows in the automation technology, robotic process automation (RPA) seeks to move these bots from desktop computers to the cloud. RPA in the cloud could simplify the infrastructure, improve scalability and provide better integration with other cloud applications.
Cloud providers are seeing the benefits of RPA and are acting accordingly. Recently, Microsoft revamped its Power Automate tooling for the cloud and Google invested in Automation Anywhere, one of the leading RPA vendors. Additionally, all the major RPA vendors have been busy refactoring their offerings to run more efficiently in the cloud.
Learn how RPA works in the cloud, its deployment model, automation workflows and the importance of governance.
How RPA works in the cloud
In many ways, RPA in the cloud works similarly to how it works on premises. IT teams create bots that can learn and then execute rules-based business processes. In addition, many RPA offerings can observe human digital actions and then design a bot to complete these actions, automating the bot creation process itself.
In the cloud, RPA can take advantage of cloud-native architectures, security models and scalability more efficiently than it can on desktop or on-premises servers. Designing the technology, design approval, security reviews and the actual establishment of RPA platforms can be a major cause of delays when kicking off a digital workforce program in a larger organization -- but the cloud option reduces those delays significantly, said Maurice Dubey, executive director at Q4 Associates and author of Adopting a Digital Workforce.
With architectures built on containerized microservices and serverless infrastructure, cloud RPA provides better scalability, said Amardeep Modi, practice director at Everest Group, an IT advisory firm.
Containers reduce the time required for configuration and setup, as well as simplify auto-scalability without manual intervention. This lowers resource utilization since IT teams can scale up and down microservices' underlying RPA capabilities independently, rather than the traditional approach of replicating the whole servers. These factors can lead to a lower total cost of ownership.
New deployment model
One important aspect of cloud RPA is the deployment model. RPA infrastructure can be dynamically scaled up in the appropriate cloud platform to be closer to other applications. This reduces the burden on IT staff that had to traditionally manage physical servers. All the RPA vendors have created cloud-specific automations that improve the provisioning and management bots on the cloud.
For example, the Blue Prism RPA platform currently supports deployments on AWS, Google Cloud Platform, Microsoft Azure, IBM, Oracle Cloud Infrastructure and Salesforce AppExchange. It also enables IT teams to integrate bots into cloud-native services, such as cognitive services and productivity applications. Another aspect of cloud support lies in tapping into an ecosystem of integration partners that specialize in different cloud platforms or business domains.
Cloud RPA changes the way RPA bots are deployed and integrated into other applications. Traditionally RPA bots have connected to other applications by clicking and scrolling in the application's user interface. The cloud makes it easier to create hybrid bots that are programmed like traditional bots but can also automate tasks via APIs.
For example, Microsoft's Power Automate calls these hybrid automation cloud flows. These allow bots to control apps via direct API integration rather than trying to mimic keystrokes and mouse clicks, which are less scalable and prone to break when the UI changes.
One of the biggest attractions of RPA, compared to low-code development tools, is that application development mirrors the way users traditionally interact with applications. This makes it easy for business users to understand what is going on and to use process mining techniques to create the starting point for new automations. APIs and low-code development tools typically require expert knowledge to use.
The rise of hybrid bots running in the cloud could combine the ease of development and understanding of RPA with the traditional performance and scalability of low-code/no-code development.
Governance is key
Cloud RPA also allows enterprises to change the way they govern automations, even for ones that automate desktop UI of an application running on a local PC. This centralizes administration and governance so that administrators get greater visibility into everything that is created and run by users in the organization.
Also, IT leaders need to ensure governance is implemented appropriately.
"By giving RPA to citizen developers, we may end up with the Microsoft Access scenario where organizations' IT departments ended up having to support a lot of poorly designed and developed Access solutions, some of which ended up managing mission-critical aspects of their business," Dubey said. He believes that baking some of the governance capabilities into cloud RPA could help reduce this risk.
Evan Kruger, manager at Deloitte Consulting, said he believes cloud RPA could also help organizations manage risk and operational support better than low-code/no-code automation. He predicted demand for cloud RPA will grow faster as more organizations move to cloud-native applications in general. This acceleration will be fueled by independent and professional developers creating and sharing automations through the cloud.