Robotic Automation and Artificial Intelligence are changing the job landscape – but in different ways

• Oxford Economics: up to 20 million manufacturing jobs worldwide will be lost to robots by 2030.
• World Economic Forum: automation will displace 75 million jobs but generate 133 million new ones worldwide by 2022.
• Gartner: AI-related job creation will reach two million net-new jobs in 2025.

It is a well-known fact that robotic automation and artificial intelligence are rapidly changing job landscape. In media, robotic automation and artificial intelligence often appear to be the same thing. However, they are quite different: the underlying technologies are different, and they are suited for distinct purposes.

RPA reduces unnecessary work steps

Robotic Process Automation (RPA) focuses on the automation of human tasks by training software to mimic human actions. Most suitable tasks to be automated this way are relatively simple repetitive actions that always follow the same clear procedures. Examples of these are:

• Sorting queries and offering initial responses to customers in Customer Service
• Automating the data input and processing certain workflows required for Invoice Processing
• Verifying employee data consistency across multiple systems, validate timesheets, load earnings, and deductions in monthly Payroll processing

RPA software’s value comes from the fact that it works at the “presentation layer.” RPA accesses and writes to applications through User Interface relatively simply following a pre-defined workflow.

This feature is handy, especially when there is a need to repeatedly load data residing in separate systems and formats, e.g., into an Excel spreadsheet. Instead of cutting-and-pasting the data manually, RPA software can be trained to do this automatically.

So, rather than embarking on an expensive integration project of back-end systems, RPA software can do data integration effortlessly in the workstation. At the same time, it removes unnecessary work steps.

RPA can be implemented in several ways depending on the scope of the deployment. It does require practice, but employees can easily adapt to the changes and work productively alongside the robot software. They can even “train” the software for new processes without expert IT-personnel.

AI is augmenting human decision making

Artificial Intelligence (AI), on the other hand, is trying to mimic human reasoning. It has three main characteristics:

• It captures information (Vision Recognition, Sound Recognition, Search, Data Analysis).
• It works out what is happening (Natural Language Processing, Reasoning, Prediction).
• It understands why something is happening (Machine Learning).

Examples of the current use of AI are:

• AI-enabled customer assistants can answer simple questions like letting the customer know the status of his order and helping him in finding a particular product based on his description.
• In health care, AI-powered technology helps pathologists in analysing tissue samples and thus, in turn, making a more accurate diagnosis.
• AI-based chatbots are changing the travel industry rapidly by facilitating human-like interaction with consumers for travel recommendations, better booking prices, and faster response times.

A common misunderstanding about AI is to focus on the pure automation of tasks. Current capabilities of AI, however, are best suited for augmenting human decision making and interactions. AI’s strength indeed is the ability to classify information and make predictions faster and at higher volumes than humans can accomplish on their own.

Algorithms are the foundation of AI. In mathematics and computer science, an algorithm is a finite sequence of clear, computer-implementable instructions, typically to solve a class of problems or to perform a computation. Effective AI application needs to utilise the right kind of algorithm.

Even though the mathematical theory behind algorithms is quite complex, software tools for creating AI applications are getting increasingly simple. They have a built-in set of algorithms, graphical workflow engines, and other solutions that make the creation of AI solutions relatively easy.

AI and RPA are complementary technologies

Transactional, knowledge-based, and strategic processes together construct all business processes. Traditionally, automation focus has been on transactional processes where also RPA is widely deployed. AI-based applications and pilots increasingly reshape knowledge-based processes. Much of transactional and knowledge-based processes will be automated. Still, there will always be a need for humans, e.g. to take care of exceptions and make decisions.

However, strategic processes will not be automated any time soon, but human experts will handle them entirely. Accordingly, Robots, Artificial Intelligence, and human employees will work side by side, all managing tasks best suited for each.

In digitalisation, AI and RPA are complementary technologies. They are used from simple desktop-based RP-automation through centrally managed “bots” to select AI-based services that assist in RPA coordinated automation. The goal over the next 3 – 5 years is AI-applications to learn processes dynamically and to make decisions accordingly.

“Once you’ve developed a rock-solid understanding of AI and its potential applications, it’s time to make a case for a pilot.” Gartner Group.