On the 18th of June, the Australian Centre for Robotic Vision (funded by the Australian Research Council) released Australia’s first Robotics Roadmap. The document aims to ‘guide how Australia can harness the benefits of the new robot economy’ and it joins a host of other international roadmaps, confirming our march toward the fourth industrial revolution.
Just four days before, our panel gave us insight into how robots are already working among us, and - with the help of a Star Wars simile - we learned that there’s more to this technology than Terminator, A.I Artificial Intelligence, or even Wall-E would have us believe...
What Robotic Process Automation (RPA) is and how it works
How and where physical robots are used
The limitations of robotics
How RPA works alongside of Natural Language Processing (NLP)
Tips for determining the best applications of robotics and how to educate your workforce
The impact of robotics on the workforce
Annie Hariharan, Senior Manager, PwC
Leigh Pullen, Executive Director, CiGen
Nicci Rossouw, CEO, Exaptec
Bots, bots, bots - everyone seems to have them but what are they and what do they do?
We explored two types of robotics;
RPA is software that is robot-like; programmed to complete standardised processes that were once completed manually by a human (often times the ‘grunt work’). These robots are wholly task-based.
Physical robots include telepresence, service and companion bots. Telepresence are (essentially) iPad’s attached to a stick, controlled by a human in another location through an application that allows them to drive the bot, be seen and communicate via the hardware.
Service bots are an extension of telepresence incorporating more applications and capabilities, while companion bots are another level of sophistication again - able to speak several languages and perform both physical and digital tasks.
So, can we expect to be working alongside Ava from Ex Machina soon?
No. While on the surface it might seem natural to integrate RPA, NLP, and AI, each are unique and (for now) mostly separate in their capabilities and design.
RPA is limited by what it is programmed to do. To introduce (another) aforementioned movie reference, RPA is like R2D2 - when asked to take a map to Luke Skywalker, R2D2 will do exactly that. It will not ask questions, or contemplate meaning. However, if you ask C-3PO, it might seek clarification and context before taking the next step. Therein lies the key difference between RPA and AI, or Machine Learning. It can mimic what humans do, it can integrate with other applications, and work with a range of data inputs, but it cannot determine meaning.
There are some physical robots that are edging closer to C-3PO territory. Hong-Kong based Hanson Robotics debuted Sophia at SXSW in 2016. Since then ‘she’ has become the first robot to receive citizenship (Saudi Arabia), and the first named Innovation Champion by the United Nations Innovation Programme. Sophia is demonstrative of the integration of robots and natural language processing (similar to that of chatbots), however does not integrate this with AI.
Where else can we see RPA and physical robots being used, and what are the benefits of this technology?
The key features of physical bots and RPAs have obvious applications across a wide range of sectors, but in Australia the uptake has been slow.
Of the physical bots, telepresence and service have been used at a corporate level in private organisations, and universities, to connect employees or fulfil roles such as reception. Companion bots are scarce, but becoming more common; finding applications in health and community services. In the aged care sector they are able to fulfil the role of an in-house, round-the-clock, aged care worker. As our population ages, and the pressure on related services becomes greater, it’s believed these bots will provide an ideal service in helping the elderly to stay in their homes for as long as possible before moving into aged-care facilities.
Financial services and oil and gas, meanwhile, are making good use of RPA. At its core, this technology is fast, efficient and accurate. It is prime for processes that are simple and standard. Those organisations that are seeking to implement RPA, often have to review, and standardise before program development can even begin and this preliminary step can result in a sometimes 30% reduction in processing, and post-implementation can sometimes see another 50% reduction.
A rental car company implemented RPA to support customer service agents, using the program to deal with approval for mechanics costs that exceeded $800. A process that would take an agent 40 minutes to complete, and potentially put a car out of action for even longer, was reduced to 3 minutes with RPA, getting cars back on the road quicker.
Other success stories for RPA, are more difficult to find. As demonstrated by the car rental company, at its best, RPA is taking care of low-value, menial tasks and freeing up humans to focus on high-value tasks. But, one can also see how bots can result in redundancies...
Are robots replacing humans?
The introduction of bots has resulted in headcount reduction, but not to the extent that it is being portrayed.
Robots cannot truly take the place of humans in all roles and responsibilities, because they are not truly artificially intelligent - they cannot (as C-3P0 does) understand context, infer, or derive meaning. They are not the task-masters, rather the task-completers, needing explicit direction in the form of programming and data modelling.
The greatest impact RPA will have will be on entry-level, or low-skilled, positions. These are the kinds of roles that teach us how to work, and how to do our jobs, to eventually move up. Think intern or recent graduate, particularly in professions like law. If these roles become the automated realm of the robot, how will future generations learn?
What should a business consider, and how should it prepare, to integrate robotics?
The reputation of robotics can elicit trepidation and, sometimes, fear among the workforce. More practically speaking, if the proper steps are not taken prior to implementation, your robotic solution can fail to deliver on the promise of its features. In order to ensure the greatest chance of success, there are steps you can and should take;
Consider, and challenge, your motivation and objectives
Headcount reduction, or cost-play, are not a compelling enough reasons to implement robotics. There are costs associated with the exploration, implementation and maintenance of this technology that will likely see any business fall short of purely financial motivation.
Integrate and plan
Linking the implementation of robotics with your broader business strategies, particularly digital, can help you overcome the trap of headcount reduction and cost-play. It can also assist in integrating this technology into your organisational structure, and help you manage maintenance as it becomes part of business-as-usual.
Standardise and simplify
As we have explored RPA, in particular, works best with processes that are standard and simple. Often those that we think are simple, actually have more variations and history than we realise. It is worth investing time and resources in early stages, to first identify the right processes, and then revise them so they are ready for automation.
The most culturally-successful introductions of RPA, and physical robots, have been where organisations have been transparent with and educated their workforce during the integration. Some organisations have humanised their robots by holding naming competitions, others have regarded them simply, but respectfully, as tools that free them up to focus on higher-value activities.
Consider the risks
RPA particularly, cannot do anything that it is not programmed to do - it cannot make independent decisions. So risk is inherited from the programmer, or those designing the work flow. To minimise potential risks, it’s recommended that organisations approach managing a bot as you would any other employee by restricting access, and enforcing data purging.
Where is this technology heading?
Uptake of this technology will only increase from now. RPA is widely used in certain industries, though the applications are universal. Similarly, physical bots are showing great promise in health and human services and events. At the South Korean Winter Olympics, 85 robots of 11 different types were deployed to various locations throughout the course of the games - they translated, provided directions, handed out water bottles, and skied the slopes. Airports are widely tipped to be the first place most of us will start to encounter bots.
The technology itself will continue to evolve with the integration of NLP and machine learning. While this hasn’t been successfully completed as yet, investment and interest suggest it will be inevitable.
Want to learn more? Take a look at the Robotics Roadmap.