Is Hyperautomation Worth The Hype?

“Traditional automation technologies have helped industrial companies achieve huge progress in becoming safer and more efficient,” says Dan Farrell, who heads Accenture UK’s Industry X technology team. “The good news is almost all industrial players recognise the power of automation and deploy it to some degree – where digital tech and investments allow.”

Unfortunately, many organisations aren’t realising automation’s full value, for a variety of reasons.

“Traditional automation is, ultimately, a process of continuous tweaking,” says Mathias Golombek, CTO at German analytics software company Exasol. “You need to continually adjust your processes, and the equipment or tools used within your business, to squeeze out extra performance or to mitigate everyday challenges. It’s trouble­shooting to improve the bottom line.”

While this is an extremely valuable undertaking, Golombek says it’s not always a business-wide process. “That’s simply due to the breadth of data, organisation knowledge, and expertise you need to execute it successfully,” he says.

Indeed, the KPMG report suggests that many of the automation implementations that have happened thus far have been piecemeal efforts that focus mainly on cutting the cost of legacy processes and reducing headcount – with, for example, siloed efforts to automate payroll, invoice processing and customer service inquiries.

“Traditional automation technologies are largely front-loaded, meaning you have to do discovery, have the subject matter expertise, and put in the time to map out the process,” says Sam Babic, senior vice president and chief innovation officer at US software firm Hyland. “It also assumes that you know where the bottlenecks are upfront and that all of your assumptions in your process design are correct. In many ways, implementing automation often feels like a waterfall versus an agile approach.”

Terry Simpson, senior solutions engineer at automation and workflow management business Nintex, explains the problem further: “Traditional automation technologies have typically been siloed into individual applications in many cases,” he says. “Organisations will purchase a software application to tackle one problem or area. Within that application the automation component is not the primary focus of the application and has limited functionality in many cases.”

What’s more, Anders Erlandsson, head of IndustryLab, Ericsson Consumer and IndustryLab Stockholm, says that even if traditional automation could be used to increase the automation level, it might not be financially sound to do so. “That’s because the cost is simply too high or the financial benefits are too limited,” he says.

Yet a shift is happening. The rise of technologies such as advanced data analytics, artificial intelligence (AI), machine learning (ML), 5G and simulation are changing the game.

“Developments in cloud and big data technologies now allow us to crunch numbers at a much faster rate and enable AI/ML-based models to gather the insights much faster and accurately,” says Mahi Inampudi, chief technology officer at US-based software firm Envoy Global. “As access to these developments become more widespread and accessible through hand-held devices, the applications of this same technology become infinite and, in turn, even more exciting.”

These evolving technologies help with the problem of “you don’t know what you don’t know” when it comes to process. “AI, for example, could help you detect signals in data that would indicate a bottleneck or some other change in the process,” says Babic. “You may even be able to look at the whole of a process such as an existing activity log and have recommendations be made to the process, potentially even in real time. 5G and 6G will increase the ability for the Internet of Things (IoT) and other edge devices to be included within the equation, potentially allowing for large amounts of data to drive automations.”

Elmgren agrees: “In order to be more flexible, the automation needs to be data-driven so it can adopt changes in product mix and production volumes,” he says. “5G can handle the volume and density of data that needs to be collected and has the low latency needed for taking fast decisions based on the data. With those capabilities, 5G creates an infrastructure [for] flexible automation.”

These advances have given way to a new type of automation. Hyperautomation – automation that is data-driven, rather than process-driven thanks to a combination of artificial intelligence, machine learning, natural language programming and predictive analytics technologies – has made its way into Gartner’s top 10 strategic technology trends for both 2020 and 2021, promising to improve efficiency, optimise processes and workflows, and lower operational costs for businesses.

“Hyperautomation is really the collection of tools and technologies that act as enablers of technologies,” says Babic. “Taken together – and ideally integrated in some fashion – they offer the ability to automate large portions of end-to-end workflows to a degree legacy automation technologies or stand-alone automation could not do otherwise.”

“Even if specific processes or decisions are already automated, businesses should use hyperautomation to go one step beyond by thinking about entire end-to-end workflows within a company, especially between departments that are, typically, not seen as connected,” adds Golombek. “For example, businesses could explore how HR processes can influence their sales revenue. The first step in doing this is by making entire workflows completely transparent so that businesses can start to properly understand what occurs at each stage. This will allow them to start measuring KPIs and, in turn, create dashboards which enable decision-makers to see the results so they can optimise their business holistically.”

Hyperautomation also facilitates the creation of digital twins. “Hyperautomation is essentially the virtual replication of an entire organisation,” says Golombek. “You’re creating a digital representation of your equipment, or services, that a computer program is able to quantify and understand. Hyperautomation requires you to do that for every element of your organisation, and then integrate them in a manner that allows these processes to respond to one another. They can change themselves to improve their functionality, and they can identify when other linked processes are going wrong or underperforming. Theoretically, they should be able to show the past, present, and possible futures of each system.

“With that in mind, to some extent, hyperautomation is digital twinning,” Golombek continues. “Companies that commit to the hyperautomation journey naturally begin to accrue the elements that facilitate digital twins: the quality and volume of data to represent their processes and equipment.”

This concept is already attracting the attention of many business leaders. “Sixty-five per cent of decision-makers across the 22 markets in our study say they expect to use digital twins in their production process within the next five years,” says Erlandsson.

The cases are compelling. For example, by creating a digital twin of a hospital, healthcare leaders can work out the impact of tweaking staffing or changing the layout of a ward, for example. They can then understand whether a tweak in one department will create a bottleneck in another without having a physical impact on patients and workers.

A physical machine like an engine or equipment is another example. “That equipment has a bunch of sensors that can sense temperature, vibrations, pressure, speed and more,” explains Simpson. “Automation would typically have processes that are triggered based on thresholds that are being monitored. Pressure gets too high, and a notification gets sent to maintenance. Vibrations become slightly out of range, and maybe a maintenance ticket is created automatically. Digital twins are a way to test and simulate concepts that would be executed in a real tangible item.”

Hyperautomation also has huge potential for the factory of the future. The combination of advanced technologies will not only help manufacturers unlock their hidden factories, but also help them meet new demands around sustainability.

“IoT, as well as 5G and even 6G, will allow for vast amounts of sensor data to be collected and analysed,” says Babic. “Hyperautomation can help detect signals in this data that would not normally be identified by a human. This would allow for optimisations of the factory, but not before leveraging the existing data and simulations to verify assumptions and optimisations.”

“The factory of the future can be optimised within three dimensions: structure, digitisation and processes,” adds Golonbek. “The hyperautomation of a factory is entirely dependent on the existing digitisation in place. This is key, as production machines need to provide relevant information about their status and usage. Hyperautomation comes into play once you incorporate all of the information a factory provides into the broader scope of an organisation – for example, capacity planning, logistics, human resources, research, sales revenue, etc.”

Another important use case can be for identifying business risks. “That’s because hyperautomation provides a holistic view of an organisation, which means businesses can develop a thorough understanding of how their company works,” says Golombek. “Through applying hyperautomation, businesses can play with scenarios to identify where their organisation is not scalable in certain areas, for example.”

It will also open up a new level of opportunities for IT professionals to become versed in how the associated tools and technologies blend to create solutions. “Some of those opportunities will probably yield new start-ups, as well as drive existing vendors to bring together these technologies in a more consumable fashion,” says Babic.

Ultimately, Inampudi believes that hyperautomation will create more job satisfaction. “Nearly every industry has experienced how challenging it is to retain and hire employees,” he says. “Aside from business productivity, hyperautomation helps employees achieve work-life balance by automating tasks and gives human resources time to focus on higher value-add activities.”

Its potential, then, is clear to see. But achieving widespread adoption of hyperautomation won’t be easy.

First, there’s a global skills shortage to address. “This is a big roadblock,” says Farrell. “The skills needed now for a new era of high-tech manufacturing look very different to a few decades ago. Companies are consistently seeking talent in engineering innovation, with technology advancing at a quicker rate than the pipeline of people to fill them.”

That’s just the start. “Since hyper­automation is really a collection of tools, technologies and even practices, the roadblocks are really in the scale of understanding the available options and bringing together those options to create solutions,” says Babic. “It is unlikely that a single vendor will provide all of these tools in a cohesive structure. This means organisations must first understand the problems they are trying to address with hyperautomation, and then navigate the landscape of tools and technologies and choose those tools that are fit to purpose. Cloud-based vendors will have to rise to the occasion to ensure these tools are available and interoperable and that businesses have the ability to seamlessly integrate these tools into their processes.”

Golombek agrees, arguing that the sheer volume of work involved in realising the full potential of hyperautomation might put many companies off. “One of the most important things is the ideology with which you approach hyperautomation,” he says. “Many businesses will have been attracted by the concept while it was at the peak of the ‘hype cycle’, only to become a bit disillusioned when it turns out to be a journey that can take years. Companies really need to consider what they want the outcome to be, and how long they’re happy to wait in order to get there. It’s also important to consider that you might not understand your business quite as well as you think you do, or that the data that comes out might challenge your concept of operations. Once you’ve got a 100 per cent accurate digital twin of your entire business, that data is going to confront you with some home truths. That requires you to put your ego aside if you want to be truly data-driven.”