This post is part of the AI for Business series
Shedding light on AI from a non-technical business/product/design perspective.
This is post #3 in our AI for Business series.
We often associate Artificial Intelligence (AI) with flashy headline-grabbing things like self-driving cars and generative chatbots. But the unsung hero of this revolution lies in the mundane. Enter process automation: the deliberate, methodical orchestration of tasks to free humans from repetitive drudgery and amplify organisational impact. While not as flashy as forms of AI that generates poems or paints pictures, process automation may quietly be one of the most transformative forces of the modern workplace.
Robotic Process Automation (RPA)
First things first, there tends not to be any actual robots in Robotic Process Automation (RPA). While the technology was in development for a period before, the term RPA came to the forefront in the early 2000s.
RPA automates high volume, repeatable tasks that would otherwise require human action (and likely a lot of time) to perform. Think data entry, regulatory compliance, fraud detection, stuff like that.
Not quite as exciting as creating an image of Pope Francis looking fresh in a white Balenciaga puffer coat, is it? But the good news is, that’s by design.
RPA is designed to target and automate some of the most mundane and repetitive computer-based tasks and processes in the workplace. By doing so, RPA boosts efficiency for organisations.
However, RPA is not AI. Without something like a deep learning neural network, regular RPA systems don’t learn as they go. If something changes in the automated task (like a field in a web form moves, for example), a standard RPA bot typically won’t be able to figure that out on its own.
To automate these more complex tasks, you may seek the help of Cognitive RPA.
Cognitive RPA
The evolution from RPA to Cognitive Robotic Process Automation (Cognitive RPA) introduces AI’s learning capabilities into the mix. These automations adapt and refine themselves over time, moving beyond rigid, rules-based operations.
By leveraging natural language processing, machine learning, and AI, Cognitive RPA can handle unstructured data like scanning invoices, analysing customer inquiries, or even predicting future trends based on historical data. This enables organisations to automate processes that were previously too complex for traditional RPA. While still relatively new, Cognitive RPA represents a leap forward in operational efficiency, offering organisations the ability to be not just faster, but smarter.
The convergence of RPA and AI has also created the conditions for the rise of AI agents. These intelligent software systems use AI and machine learning algorithms to make decisions and interact with users or systems. AI agents can handle even more complex processes that require cognition, learning, and problem-solving.
The Value Equation
Automation for automation’s sake is a trap as old as the technology itself. Claude Shannon’s “useless machine,” which exists solely to turn itself off, humorously encapsulates this pitfall. To avoid this, organisations must evaluate the value of their processes before automating them. The key questions to ask include:
- Does this improve customer outcomes?
- Does it streamline operations without overcomplicating them?
- Will it scale effectively as we grow?
You may have seen Ocado’s robotic grocery packing warehouse. Rather than creating a humanoid robot to wander the isles of a physical supermarket, filling a shopping basket with tomatoes like some kind of domesticated terminator, Ocado ripped up their processes in favour of something totally new. Their grid-based automation system is AI-powered and optimised for efficiency. At first glance you think “Wow, this is the way of the future!”.
And that may be true, in time. But, for now, here’s the catch: consumers choose deliveries from a particular supermarket because they have better produce, they are cheaper or they have faster delivery. They don’t really care about the specific technology enabling the supermarket. The lesson? Always tether automation efforts to tangible benefits for end-users.
Make your processes easier to automate
If automation is the goal, it might be worth looking at your processes. How easy are they to automate?
Let’s take a look at the car manufacturing industry. Companies like Tesla and BMW are known for their automated and robotic car assembly processes. But even Tesla’s CEO has admitted that some aspects of their production are difficult to automate, stating that “General assembly is really well suited to people because you have got a lot of weird things you’ve got to put together.”
Rather than trying to automate a tricky task like those of the general assembly line, is it possible to replace the task with one that is easier to automate? Do you really need that part? Could you design your way out of having to do these tasks?
The humble drink can
Did you ever wonder why cans are cylindrical? Couldn’t they be more efficiently packed if they were cuboid in shape? Well, cylinders are preferred for a host of reasons, but a really core one is that cylindrical cans are easy to manufacture.
While not as efficient to pack, cylindrical cans are formed from a single aluminium disk. This process is simple, fast, and easy to automate. By contrast, pressing and forming sharp corners in metal is more difficult, leading to stress concentrations that make the container prone to deformation or failure.
We’re sure you have also lost sleep pondering the purpose of the dome indentation at the base of your cylindrical can. The dome is formed when a machine presses the bottom of the cylinder against a small semi-sphere. The strength of the dome means less metal is required than if the bottom of the can were flat. The dome can be created in a single, efficient movement in a matter of split seconds. Strengthening a flat bottom would take more time.
If we have piqued your interest in aluminium can manufacturing, here is a fantastic video on the subject.
Key takeaway? To bring your automation to the next level, consider changing your product to eliminate those “weird” things that are difficult to automate.
Zoom out once in a while
While it’s great to start streamlining your processes with automation, it can also be beneficial to zoom out once in a while.
We looked at how Tesla is automating the way it builds cars. Through affordable cars with full self driving capabilities, Tesla plans to revolutionise the car industry.
But what if we zoomed out for a moment. Cars were created to solve the problem of efficiently moving people from A to B. They were the best solution available with the technology at the time they were invented. However, when we think about the original problem of moving people efficiently, would the best solution be to create a self-driving car that looks like a car and drives like a human?
Or, would we be better off trying to create autonomous passenger drones, which would travel as the crow flies and not get stuck in traffic? Or maybe there is something else completely different and waiting to be invented?
Obviously regulations, ethical considerations and available technology are all limiting factors on this kind of big vision thinking. But, when considering automation, it can be useful to ask yourself whether a process is worth automating, or whether it should be scrapped completely in favour of a totally new, alternative approach.
To automate or not to automate?:
And now we have reached the critical question; to automate or not to automate?
Well, as we discussed above and in the previous blog in this series, it is important that you use any technology tool, including automation, to provide value to customers and not for the sake of using the technology itself.
If you are considering using RPA or Cognitive RPA in your processes, remember that in most cases, customers don’t care whether a package was packed and transported by a human or a robot; they care about their package arriving quickly and at a cheaper price point.
If cognitive automation can help you to deliver these kinds of benefits to your customers then great, it sounds like a good use of resources! If these benefits can be achieved with regular RPA, even better! But make sure automation is adding value to your customers and helping you reach your strategic goals.
💡 If you’re ready to start using AI to transform your business, thoughtbot would love to work with you. Let’s talk about making your AI initiative a success!
This blog post is part of a series based on the course Artificial Intelligence (AI) for Business run by University College Dublin (UCD). I took this course and found it so helpful that I’m recapping my top insights. thoughbot has no affiliation with UCD.
