Digital Transformation: Intelligent Automation, Robotics and RPA for the Non-Technologist, Part 2
Article 32: Understanding What Technologies Do and How They Can Be Used
Automation technologies are one of the fundamental drivers of Digital Transformation. By streamlining activities, automation can change cost, timelines, and what is possible. But just because robots and automation can be game-changing, doesn’t mean they automatically are. Using a robot to replace a human as the one rubbing clothing on a washboard to get it clean is bad Digital Transformation. Automating the effort by agitating clothing to remove dirt with a washing machine (that does not have two arms and a rough board) is great Digital Transformation. You haven’t just made a replacement. You’ve designed a better solution for the task.
Today we’ll talk more about making that shift. We’ll discuss how to move:
Keep in mind that there is nothing inherent in automation that ensures that you automate to transform.
Simplified Technical Description
The goal of robotics, RPA and intelligent automation is to help take tasks that are unsafe or undesirable for humans to do and instead do them by machine. These categories are a combination of hardware (robotics) and software (robotics, RPA and intelligent automation.) These technologies more efficiently or effectively do tasks that were done less efficiently or were impossible to do to begin with.
Robotics
The field of robotics focuses on building machines that can perform (increasingly) complex tasks. This includes doing everything from dangerous work, like spraying toxic chemicals and defusing bombs, to boring repetitive tasks, like vacuuming or assembly-line work.
Robots need to be both smart enough to know what to do and structured properly to do it. To make that happen, robotics blends multiple areas of expertise, including mechanical engineering, electrical engineering, and computer science. The machines are built so that they can take input from a sensor and then process that information as instruction, which is followed by an action. Key components to make this possible include:
Physical housing: Robotics ties the automation (or instructions) to the physical components doing the task(s). It is the only one of the three with this physical component. And an underwater exploration robot will have a very different physical structure from a medical robot that performs surgeries.
Sensors: Robots need input. Humans take in a variety of information with their five senses. Similarly, robots have cameras for vision, accelerometers to detect motion, proximity sensors for distance, input channels for data feeds, and many other sources of input.
Control Systems: The robot takes all inputs from sensors and, depending on the results of its algorithms, performs specific actions or tasks.
Power Supplies/Actuators: Once the robot decides what to do, it needs the power and capability to do it. The actuators will take the power and convert it to action given the purposes of the robot (e.g., turn a wheel, move an arm, etc.)
Many people think of something like the Terminator when they think of robots. This can lead to significant confusion between artificial intelligence (AI) and robotics. A robot may (and increasingly does) leverage AI to more effectively do its job, but that does not mean that it has to or that they are the same thing.
Robotic Process Automation (RPA)
RPA is essentially a digital robot. You use software to create a bot which (much like a physical robot) can automate the repetitive or rule-based tasks. These bots can perform the same functions as humans by using software to enter data and click buttons.
Most software available today is not fully optimized for the different businesses that run it. So, companies looking to enhance their off-the-shelf software can teach their employees (or their bots) to optimize their business. Companies can optimize by building a custom system for a differentiated part of the business. They won’t need bots for that, though they will want to tie it into other off-the-shelf software. Any place where a human does a repetitive virtual task (and some non-virtual tasks), an RPA bot can be built to digitize it.
Clever employees have been finding shortcuts to make their jobs easier forever. RPA allows you to codify these shortcuts so that it’s not just your best employees that are optimizing with shortcuts–-all of your bots are. Every process can be automated to maximize efficiency and, over time, you can evolve and adapt the bots as you find new and better ways to get to your end goal.
You can use RPA to support customers with automation of account opening or customer service responses. You can also use it for internal activities, such as order processing or new employee onboarding. And it’s not just the business teams that can use it, RPA is great for data migration and report processing. We can make every single part of a business more efficient with automation. The question is merely which parts are worth the investment.
Intelligent Automation
AI isn’t just for robots. In the RPA overview, we discussed automating the repetitive, rule-based tasks. Intelligent automation is about automating in situations where there is more uncertainty. Invest here when you need your bots to figure out dynamically what the right action is.
Using AI and machine learning (ML), you can automate to allow learning and adaptation inside of the system, without any intervention. You don’t need to wait for a human to realize something can be done more efficiently or effectively. The bot can figure it out itself. It can handle the complexity and variability that are part of so many processes and find the shortest or best path to the goal.
Intelligent automation allows increasingly onerous tasks to be automated. As AI and machine learning advance, the ability to use these tools to fully optimize every process will follow. The biggest consideration here involves responsible use. As we build these AI and ML models, we need to make sure that we are building them ethically and responsibly. There are already too many examples of bias and lack of responsible usage. Once we embed these systems as bots into other systems, we constrain our ability to see problems. We need to be incredibly careful that we are not creating systemic, invisible problems.
The examples of applications for intelligent automation are just as endless as with robotics and RPA. But rather than the rules-based processes you find with RPA, intelligent automation can handle unpredictability. In customer service, we automate with chatbots or by creating personalized shopping experiences. In healthcare and financial services, we automate issue identification to help with diagnoses and fraud discovery, respectively. We also optimize processes that change over time, like delivery routes or manufacturing quality. Any dynamic process can benefit from this combination of basic automation with the intelligence of AI and ML.
Summary
Intelligent automation, robotics and RPA are two continuums. One is a physical automation where you have robots that are anywhere from entirely rules-based to smart using AI and ML. The second continuum is the digital or software-based bots, where again you have the option to go from fully rules-based to more intelligent.
The best choice of robot or bot is entirely based on your goal and your business case. Technology is constantly evolving to extend what is possible in making robots and bots intelligent. And this extension will continue to bring the costs of the entire continuum down.
If you missed Part 1 in this series, click here. And if you want Part 3 delivered right to your inbox, sign up here to subscribe.