There’s no question that robotic process automation technology provides a number of straightforward and powerful business benefits. It can relieve people of mind-numbing repetitive tasks, cut costs, and increase quality. But can it also deliver a more substantial benefit? Some say yes: “RPA provides the benefits of process automation AND advanced operational and process analytics”; “Robotic process automation will radically simplify entire areas of the CIO's province, including customer analytics, data mining, social media analysis and the warehousing of big data.” Is this really true?
Different Types of Analytics
Before that question can be answered, let’s look at different ways the term analytics is generally used in a business context.
Historical Data: Often referred to as Big Data, historical data analytics focus on modeling existing data (often terabytes or petabytes of data sets) into meaningful management information for tactical and/or strategic decision-making.
Monitoring & Control: This type uses computing networks to collect and model operational data to guide the performance of related technology. An example would be energy companies that must engage with the national power grid as it dynamically moves electricity around the country, seeking to efficiently balance supply between producers and consumers.
Streaming Data: In contrast with Big Data, streaming data immediately models available data to support real-time management decision-making. An example would be an airline continuously monitoring and modeling its competitor’s web-based pricing data.
Its clear robotic software does not provide either Big Data or Monitoring & Control value. At first glance streaming data appears to be a possibility. After all, the software can easily be configured to gather web-based pricing information from competitor websites.
Where the streaming data possibility falls apart is on the word ‘models’. In each of the three types there were differences in the nature and timing of data. However, all of them applied existing data models to create information that would support management decision-making.
RPA & Data Federation
Let’s look closely at the two quotes cited in the beginning of the article.
“RPA provides the benefits of process automation AND advanced operational and process analytics” Examining the linked article, there is no reference to any type of modeling capabilities. Rather, the analytic benefits are tied to the technology’s use of “existing user interfaces to client's enterprise applications (both in house and in the cloud) and does not require back end integration or access to client data through API's.”
“Robotic process automation will radically simplify ……. analytics, data mining, social media analysis and the warehousing of big data.” Again, checking the linked article finds no mention of modeling capabilities – only a repetition of automation benefits, “As companies transition to [RPA] …… they will aggressively re-engineer themselves…and execute sophisticated and targeted data analytics with increased speed, quality and scalability.”
There simply isn’t any evidence to support the contention robotic process automation has analytic functionality. It does provide strong capabilities to support Streaming Data by dynamically collecting data across the web and from a myriad of internal sources and databases. It also has excellent capabilities to support Historical Data with efficient data cleansing and migration features.
What can be said is that robotic software does provide the value of data federation – the capability to collect data from many different sources and aggregate it in a source and format that can be easily used by existing business intelligence (BI) or other types of modeling analysis.
In fact, the relationship between data streaming analytics – which requires a robust data federation, and robotic process automation – which provides it, is so close it could be accurately termed a partnership. Rather than perpetuating a shaky claim to a analysis capability, a real service to customers would be done by letting it go and placing a spotlight on the data federation performance RPA does so very well.