It is a truth, universally acknowledged, that the claims made by marketing departments of what their technology can do are always months, if not years, ahead of what the technology can actually do. This inevitably leads to ‘marketing inflation’ as each vendor tries to show that they are further ahead of the curve than their competitors.
And so it is with Robotic Process Automation (RPA). For a few years everyone was happy with calling RPA just RPA. But then, as more players came on to the market and competition heated up, some vendors sought to outdo their rivals by bolting on some basic Artificial Intelligence (AI) functionality that would allow their software to cope with, for example, unstructured data inputs. This now meant that they could claim that their software was ‘intelligent’.
At this point we need to take a step back and remind ourselves what it is about RPA that makes it so useful. One of the key benefits of RPA is that, once trained, the robots will do that process in exactly the same way again and again and again. For this to happen the robots must be ‘dumb’ – they cannot show any degree of autonomy or self-learning, otherwise they might start doing something different than what you want them to. To use the scientific term, they are deterministic machines.
Artificial Intelligence, on the other hand, is probabilistic. It (for the most part) generalises data that it has been trained on to find the most probable answer or outcome. Which is great when there is variability in the format of incoming data (which robots can’t handle) or there are complex decisions to be made (again, something that RPA can’t handle). So, a combination of dumb robots and self-learning AI provides us with the tools to automate a wide range of business processes. But is it yet Intelligent Automation?
The marketing teams would, clearly, claim that it is. But, in our world-view, Intelligent Automation (IA) is much more than a few technologies bolted together. Of course, it has to be more than RPA-at-scale, so would have to include some significant AI capabilities. It could also include other disruptive technologies such as the Internet of Things (IoT) or Distributive Ledger Technologies (DLTs, sometimes referred to as Blockchain).
But, for us, it is more about the scope and scale of the automation than it is about the technologies. The thing that makes Intelligent Automation different is in the ambition: it is not simply automating a few processes with some RPA and AI, it is about transforming a function or whole business using an array of disruptive technologies and new ways of working. For example, in an HR department, we could automate the employee onboarding process using RPA to handle the repetitive elements and some AI to extract data from the new employee’s verification documents, which would certainly bring some useful benefits.
But, with an IA approach, we would be looking at how we can transform the way that the department engages with employees. For example: automating the appropriate interactions; predicting employees needs so that they can be proactively served; providing a 24×7 service for priority transactions; being able to explain complex HR processes in simple terms; and being able to respond immediately whenever or wherever the employees might need information. To do this would probably require RPA, chatbots, predictive analytics, dynamic web forms, personalised portals, voice recognition, facial recognition, mobile apps and process reengineering. But, taken together, these approaches will give the employees the best possible service whilst maintaining an efficient and cost-effective HR department. But is it yet Intelligent Automation?
Well, almost. You see, the real benefit of this approach will be the time the HR teams now have available to spend with employees on those really difficult, complex and/or sensitive issues. By first defining what those moments that matter are, such as when an employee comes up to retirement age, or they have a really challenging staff member, or they have some tricky problem at home that means they need some time off, then everything else beyond that just becomes a transaction that could be automated.
The intelligence in Intelligent Automation is understanding which parts of the business still need real people to be involved. The intelligence, therefore, is human, not artificial.