Robotic Process Automation (RPA) is experiencing rapid growth due to its capacity to automate manual and repetitive rule-based tasks such as accounts payable and claim processing previously performed by humans. As the RPA market has accelerated, the tools have started to mature and we see an opportunity for RPA to play a more strategic role in the integration digital ecosystem (a concept discussed earlier by my colleague Riaan Ingram) as well as providing the foundation for future Artificial Intelligence and Cognitive Computing technologies.
The Data Integration RPA Use Case
Despite all the modernization of technology in recent years, many enterprise businesses still have legacy systems in their environment. You know the type, 30-40 years old, mission critical, with no ‘pipes’ to transfer data in or out.
Automating data integration with these systems is a common business challenge that can’t always be solved in a cost effective way with an iPaaS, API Gateway or ESB. Some businesses will automate integration with these systems intrusively via the ‘back door’ but this is time intensive and complex technology architectures must be built to support this.
What happens more commonly is that humans transfer data manually in and out of these systems and this is where RPA can play a significant role. RPA allows data transfer through the ‘front door’ which can be easier and faster to implement than via the back door.
We were able to test this theory recently when of one of our clients was looking to implement a modern Human Capital Management (HCM) system. To meet business needs and maximize value the HCM will need to be integrated with the Payroll system, however their Payroll system is 20 years old and has no API for integration. For a specific use case we trialled two approaches for integration, a custom built API and an RPA tool. Both options were viable however in this instance the RPA solution was approximately 50% faster to implement.
RPA can also be a great short-term tactical option for integration requirements whilst the long-term solution with full logic is developed. For example, one of our clients needed to respond to a legislation change that required the amendment of 6 data fields. This could have been done with extensive manual data entry effort, however, the opportunity was taken to invest some of this effort into RPA bot development, delivering a fast tactical outcome with higher compliance and data quality, and ultimately a reusable asset for future events.
RPA Best Practice for Data Integration
As with any technology RPA has its limitations, and we are not suggesting it is the best option for all data integration scenarios. RPA is probably not a cost effective solution for high transaction volumes e.g. millions of transactions a day. It is also probably not a good fit for situations where every transaction or the majority of transactions are a ‘variation’ of the base flow.
In an enterprise environment, it is important to apply fundamental software principles such as auditing, logging and security to your RPA solution. RPA can be quick and easy to implement which is great, but a bot can do much more damage than a human if an error is made. You should also manage your RPA solution in a product management lifecycle (version control). As systems are updated the User Interface (UI) may change so the RPA bot will need to be updated.
Whilst the RPA bots may replace some human activity it does not replace the need to have performance expectations on that activity. To maintain a healthy process, KPI’s for the bot service need to be set and measured against to ensure they are performing their role.
Change Management
Traditional back end integration happens in a ‘black box’ that is not visible to the business. Integration with RPA happens with bots at the front end, they have personas that the humans in your business can see and interact with. You need to define how the bots will co-exist with the humans in your environment to ensure successful adoption.
The Last 20%
APIs are always the best approach for data-based integration. However where an organisation has a legacy system, they would always hit a limit to their automated integration capability no matter how sophisticated their integration platform. Modern RPA technology offers a genuine alternative for businesses that want to automate the last 20% of their integration scenarios.
