RPA (Robotic Process Automation) has been getting quite a lot of coverage lately. The idea that certain things can be automated to remove human error from the equation is nothing new, and different ways of providing such automation have been around for many years. RPA nominally takes things a little further—it can copy and codify an individual’s actions, meaning that little to no knowledge of scripting or coding should be required.
Sounds promising, but this technology only goes a certain way toward solving an organization’s problems, particularly when it comes to data.
RPA is particularly good at carrying out fixed activities with simple triggers—such as moving a set of data from a known point A to a known point B at a fixed time. What it isn’t very good at doing is moving data from a known point A to any of a set of points, depending on context or based on variable trigger criteria.
This is where workload automation (WLA) comes in.
The Institute of Electrical and Electronics Engineers (IEEE) has attempted to create a standard means of using technology phrases—not that the vendor community seem to be paying that much attention to what these are. However, if we look at how the IEEE defines RPA, it says that it should be used to describe a “preconfigured software instance that uses business rules and predefined activity choreography to complete the autonomous execution of a combination of processes, activities, transactions, and tasks in one or more unrelated software systems to deliver a result or service with human exception management.”
Note here that this definition still includes ‘human exception management’ as a key aspect: it is expected that humans will still play a large roll in RPA.
In the case of artificial intelligence (AI), the IEEE similarly defines it as being “the combination of cognitive automation, machine learning (ML), reasoning, hypothesis generation and analysis, natural language processing and intentional algorithm mutation producing insights and analytics at or above human capability.”
This definition, on the other hand, aims at removing the human aspect as much as possible. Bear in mind that several research studies have shown that around 80% of problems within an operational IT environment are caused by humans; that is, it is the carbon-based lifeforms that cause the problems, not the silicon-based ones.
The idea with WLA is to harness as much data and operational intelligence as possible to effect desired outcomes with less input from humans. The final target is what is known as idempotency: the capability to define the required end result and leave the system to its own devices to get there—no matter what. Increasingly, this is where WLA is going: it is capable of defining target storage areas for data based on business-defined needs such as performance, costs, availability, etc. It has the ability to choose the right time for an action to take place, whether this be according to cost, bandwidth optimization, immediacy, or some other factor. On its own, RPA cannot do this.
Does this mean that RPA is a busted flush? Not at all—RPA tends to be a resource-efficient means of carrying out relatively simple actions that need carrying out time after time after time. Once it knows how to do something, it will replicate that action without any problems at all. Bear in mind, however, that RPA does exactly what it is told; that is, if it is fed garbage, it will replicate that garbage until it is stopped. Think of what happened when a server configuration change cascaded out of control for Google Cloud Platform in July 2019, causing major shutdowns not just of its own environment, but across the whole internet as traffic spikes caused traffic management issues around the globe.
Intelligent RPA would have seen the problem earlier on and stopped the cascade from happening. As it was, it just kept on doing what it had been told, causing major problems in the process.
However, WLA can make use of RPA: rather than using an intelligent engine to come up with the same decision every time, WLA can be used as a surrogate human that feeds the required data to an RPA environment for tasks to be carried out. This leaves more resources for what WLA does best: the more complex stuff that needs more thought.
WLA is thus the ideal enabler for RPA. When RPA's decision making combines with WLA's robust execution functions, security and speed, the result is a dynamic suite of automation using the best and latest advancements in IT.
Also, the WLA engine can monitor what the outcome of RPA activities are, and act to trigger events that can shut down what the RPA system is doing, to try to ameliorate the impact or to alert a human if that is required.
Basically, RPA still has a part to play, as long as it is used for the right tasks, in the right way, with the right monitoring and controls in place. WLA also has a major part to play in managing data flows across a highly dynamic virtual environment, becoming a necessity where dynamic hybrid clouds are concerned. WLA can, and should, use RPA where it makes sense, but cannot just hand off its requirements and forget about them. WLA must be capable of monitoring the end results to ensure that what was expected to happen did happen—and to take remedial actions if this is not the case.
In short: with the right automation strategy in place, RPA and WLA will synergize seamlessly to become part of a dynamic automation roadmap moving your company through the digital transformation landscape.