Process, data and workflow automation

Computers help us in so many ways, but can they do more to make our lives easier, with greater accuracy and safety online? In the first of two articles, we look into the differences between automation and artificial intelligence.

Robotic process automation (RPA) - what is it?

RPA is a technology that makes it easy to create software programmes – known as robots, or ’bots‘–  that mimic the way humans interact with computers. Bots can be set up to perform high volume, repeatable tasks such as logging into applications to enter data, calculating and completing tasks, and copying data between applications or workflows. By doing this, humans can delegate boring, time consuming tasks to the machines and use their time more effectively for more stimulating tasks. 

RPA can be used in isolation or it can be combined with artificial intelligence (AI) and machine learning to do even more clever things. For instance, RPA can be used in combination with a technology called ’Optical Character Recognition‘ (OCR) to read text or handwriting. By doing this, a bot can extract data such as names, or addresses from invoices or other similar standard documents to make it much easier and faster to process them, and this will be of help to finance teams.  

In the sections below, we explore the differences between automation and AI and the benefits, opportunities and challenges associated with this emerging technology. 

For background information, watch this YouTube video: What is RPA? ​ 

Automation or AI – How are the​​​y different?

At a basic level, RPA is associated with ‘doing’, whereas AI and machine learning are most associated with ‘thinking’ or ‘learning’. 

If we take invoice processing as an example, we can use RPA to perform repetitive tasks where there is no learning involved, for example: retrieving emails, downloading attachments into a specified folder or creating bills. Here, an RPA programme can be set up with clear instructions to collect data and process information. The bot itself does not understand the data; there is no learning involved and no interaction with humans. This is all about executing tasks, processing structured data and performing actions that are repetitive - such as cut and paste actions, logging into applications or connecting to application programming interfaces (APIs) to save us time. 

On the other hand, AI and machine learning are technologies that can go from ‘execution’ into ‘thinking’. AI and Machine Learning can be trained using large amounts of data. They can learn to recognise patterns in the data sets. In this way, AI and machine learning algorithms can learn to make decisions and handle complex processes without human intervention. They can understand documents, comprehend conversations, process language and interact with humans. AI and machine learning can also process unstructured data, learn and evolve. They can decide on their own without human input and perform non-repetitive tasks. 

This table shows a summary of the main differences between RPA and AI: 

Robotic process ​automation AI and machine learning
Repetitive tasks ​Non-repetitive tasks
​No learning Learning and evolving​
​No interaction with humans ​Interacts with humans
​Collects data but doesn't understand it Understands data
Follows instructions 'Thinks' on its own
​Structured data Unstructured data
​Execution ​​'Thinking'/decision-making


Learn more ​

In the next article we will discuss benefits, opportunities and challenges associated with this close coupling of humans and technology. In the meantime, if you want to find out more about RPA, or you would like to discuss ideas about possible applications of RPA in your area of work, please email

Background reading

Robotic Process Automation: Contemporary themes and challenges

Video on RPA from an IBM perspective 

​What is RPA (Robotic Process Automation)?​ ​ (opens in YouTube)

IT Help website page

Microsoft Power Automate help page