The rise of Robotic Process Automation (RPA) has been rapid, and it is expected to continue. Deloitte’s third annual RPA survey predicted that RPA adoption will increase from 53 per cent to 72 per cent within the next two years. Yet, while a lot is known about RPA’s capability to automate mundane data heavy processes, little is said about its ability to act as an integration tool.
Both RPA and an Application Program Interface (API) effectively provide a channel of communication to move information between two or more separate pieces of software. This makes them both useful tools for enterprise application integration (EAI). But when should organisations use RPA and when should they use an API to integrate their systems?
In its simplest terms a piece of RPA software takes over the mundane interactions of an existing application. A human no longer has to click buttons, cut and paste values, or type data into fields. In short it automates repetitive, rules-based processes usually performed by people sitting in front of computers. By interacting with applications just as a human would, RPA can open email attachments, complete e-forms, record and re-key data and perform other tasks that mimic human action.
The benefits of RPA
Within this context RPA can be useful when the interactions are with older, legacy applications. It can breathe new life into legacy systems and create digital process flows, where before there was only spaghetti code, manual workarounds and swamps of data. In contrast, developing an API is expensive in terms of development time, ongoing maintenance, providing API documentation and providing support to users of your API.
RPA offers a far cheaper and quicker solution to the same problem that an API typically looks to solve. RPA’s competitive advantage over an API is its light structure; it offers virtually a no-code alternative to an API. This can provide huge advantages for businesses looking to digitally transform as it means that they can implement new technologies without having to make changes to their existing IT systems. API’s on the other hand can be limiting for system heavy organisations in this respect as implementing and integrating new systems often requires them to have to reinvent their existing IT.
Rather than replace the old, RPA acts more like a human worker would, creating a communication bridge between the two separate IT systems. A software robot can be taught a simple process in just a few days, which means that organisations can realise the benefits of the integration quickly. The same software robots can also be quickly reconfigured to integrate different systems.
In addition, using RPA to integrate two systems requires much less skill, whereas implementing an API would normally require a specialised systems engineer to configure. Of course, this requires significant investment and to some extent rewrites the rules for existing IT systems. But RPA is often limited when it comes to integrating more nuanced and complex systems and this is where an API may be of better use.
The retail example
To illustrate the differences let’s look at RPA and API in retail. RPA eases the most basic tasks such as product scanning, data analytics, and inventory management by linking systems and processes and performing repetitive tasks. In e-commerce, for instance, the return of goods is typically a manual time consuming task which increases costs. However, with the use of RPA the management of returns and the required system changes such as altering the inventory database and customers’ billing are automated.
But if the same retailer wants to dive into its data pools, pull out reams of information, slice and dice and feed new data between applications to identify more selling opportunities an API is more appropriate. Within this context an API is essentially the code that governs data access points for a server, or servers. It acts as an access broker to different applications and provides functionality and data access across different applications, which can include business intelligence and data analysis.
In short, APIs integrate across applications as 'digital' plumbing, allowing programs to interact with each other. They are expensive and time-consuming to develop but in a sense they run under the ground enabling deep rooted system integration changes. As such they also require advanced planning to manage using API management software. RPA on the other hand runs above the ground to automate existing rules-based processes such as automating inventory database entries when goods are returned. An RPA is ideal for making quick changes by creating a software robot that mimics human actions to complete a specific task.
Ultimately, there are pros and cons to using both RPAs and APIs. IT departments will require a plethora of different tools in order to optimise their organisation’s IT environment and this means evaluating what tool will best fit the integration. There’s a time and place for everything and different solutions work in different situations. System integration is great, when the solution needs to be robust. But in many cases, RPA is enough where full system integration would be overkill.
Niko Lehtonen, Head of Training, Digital Workforce