Every year, Gartner gives a rundown of what it predicts will be the major strategic trends for companies to explore in the coming year, and 2019 was no different. The leading research and advisory firm has some lofty expectations for the year ahead—including practical blockchain applications, an increase in distributed cloud computing, more focus on transparency, and the democratization of expertise, among others—but one of their new strategic trends in particular caught our eye at SEGRON: hyperautomation.
As automation aficionados, this is right up our alley. At the same time, its inclusion on Gartner’s list might make you wonder, “What implications might this kind of strategic initiative have for telecoms in general, and telecom testers in particular?”
What Is Hyperautomation?
Okay, first things first, what is hyperautomation, and how is it different from other forms of automation? According to Gartner: “(H)yperautomation today involves a combination of tools, including robotic process automation (RPA), intelligent business management software (iBPMS) and AI, with a goal of increasingly AI-driven decision making.” The operative concept here is augmenting human abilities so that they can work more efficiently and arrive at the correct decision more quickly. This is being powered by the democratization of testcase development and test result processing.
In this sense, hyperautomation isn’t really a new idea—rather, it represents the integration of different ideas that have all been at play to varying degrees in various industries for some time now. Telco operators might think of it as a sort of vehicle for technological synergy, bringing a new level of sophistication to more process-oriented automation.
Thus, the real difference between hyperautomation and existing automation technology like RPA is that it layers AI and machine learning workflows on top of and alongside what’s already there. The result is, ostensibly, the potential for new optimizations in any and all use cases that currently fall under the umbrella of either process automation or artificial intelligence. The effect should be to speed up digital transformation and make automation easier to implement (and utilize) at every touchpoint on the org chart.
Hyperautomation Use Cases
Because this is a new and emerging concept, it’s hard to predict exactly what it will look like in practice. That said, based on our own experiences with automation, we foresee a few scenarios.
Test scripts: At this very moment, things like Robot Framework are using automation to improve the overall reusability of test scripts. Rather than crafting a new script from scratch each time you need to test a new device or network element, you can substitute pre-defined keywords in and out and execute tests in a modular fashion. This already presents a huge potential for ROI when it comes to performing, say, regression test suites that need to be adjusted and run with some frequency. In a hyperautomation scenario, you might integrate AI technology into the existing level of automation in order to automatically improve, refine, and adapt your test scripts over time. This might reduce some of the upfront manual effort that people associate with keyword-based tests, and also save some of the (already modest) time and effort that goes into keeping those test scripts up-to-date. By the same token, if there are any redundancies in the underlying scripts, the AI layer could identify them and root them out.
Network equipment: Right now, telco operators are beginning to leverage usage data from various network equipment in order to schedule maintenance before a device or element breaks down. This has some obvious benefits for operators. Specifically, it has the potential to reduce network downtime by staving off unexpected equipment breakdowns. With hyperautomation, these kinds of predictive analytics can be made even more powerful by combing them with process-oriented automation. Operators might, for instance, implement a system that not only identifies predictive maintenance opportunities, but also automatically makes adjustments to OSS/BSS systems to account for the downtime.
Signalling trace: Moving beyond end-to-end tests to automatically capture information about the system under test is already possible in SEGRON’s current ATF technology. Telco testers can gain valuable data about which operations and procedures (e.g. tracking area update, SIP messaging, NP inquiry) are in execution in the network while a testcase is running. This can make a big difference when looking for root causes of bugs and other gaps in service. With hyperautomation, this kind of workflow might be expanded using machine learning to help correlate likely bugs with different trace captures. In this way, testers would be able to go beyond hints and clues about root causes to actually pinpoint issues in a data-driven way.
These are just a few examples to illustrate what we think this might mean for testers in the telco domain. No doubt, some of the most exciting applications for hyperautomation are things that haven’t even been considered yet.
What Is a Digital Twin?
Towards the end of its section hyperautomation, Gartner makes note of another potentially impactful technology: “Although not the main goal, hyperautomation often results in the creation of a digital twin of the organization (DTO).” As with prescriptive maintenance, DTOs have mostly been employed in a manufacturing capacity to this point—but they do have the potential to make an impact for telco operators. How? By providing a virtual model of one’s network for the purpose of testing out potential changes in a low stakes environment.
Obviously, this would never be a substitute for actual testing on real, out-of-the-box devices in the event that you did make a change to your network or rolled out a new service offering. At the same time, this kind of technology can help you to lighten testers’ eventual loads by figuring out some of the finer details before you make live updates. Like the rest of the innovations we discussed above, this requires integration of both AI or ML on top of any existing automation, but the potential strategic benefits can be significant. Testing out new ideas on a virtual model of your network can help reduce outages, speed up new service rollouts, and, ultimately, have a positive impact on your bottom line.