Defining automation, bots & artificial intelligence


Defining automation, bots & artificial intelligence

There is always something being hyped in tech, about a decade ago it was cloud and more recently it is multi-cloud. Right now, we have a dearth of technologies being thrown at us that include, artificial intelligence (AI), blockchain, robotic process automation (RPA), bots, the list goes on.

The trouble is that vendors and suppliers are in a constant battle to look relevant to their customers. So, they look to find ways to talk about these technologies to demonstrate relevancy. I don’t have a problem with an industry extolling the virtues of the latest and greatest but my issue is that terms are taken out of context and, at worst, just used incorrectly.

Cloud clarity
Some time ago, I was speaking to a storage vendor who had launched their ‘Cloud Storage’. I was surprised to hear that their product would be shipped to a customer’s site, connected to their servers and the customer would be responsible for looking after them. What had changed was they way they paid for it, which was now in line with how they consumed it. In no-way was this cloud but the customers thought they had ticked the cloud box.

That is an extreme example but I feel it is happening again with new technologies. Companies working with blockchain technology are some of the worst offenders, the others are around AI and machine learning.

What is automation?
One area that seems to be suffering from confusion are automation technologies. The words connected with this field include, RPA, bots and AI. These can all be used as individual components to automate processes and or be brought together but they are not the same thing.

These are some of the combinations I see that enforce this confusion:

Bots and Robotic Process Automation (RPA)

Robots and Artificial Intelligence (AI)

AI and Machine Learning

It’s good to talk
If we take the term bots shortened from robots, it is usually sat alongside chat so we get the chat bot. They are a technology solution used to mimic a human having a real time messaging discussion for the delivery of a service. Chat bots automate simple processes to manage peak demand without the need for additional staff. They are used in conjunction with machine learning technologies that process natural language to make the conversation feel ‘real’.

Defining Robotic Process Automation
Robotic Process Automation (RPA) is nothing like a chat bot. This technology will mimic a human and it does so within the confines of a specific process without human interaction. It is deployed to automate highly repetitive tasks that use pieces of software that cannot be integrated via application programming interfaces (APIs), in these scenarios it is the human that provides the interface between multiple pieces of software.

For example, imagine changing a person’s address within a local authority, that person may exist in a number of different systems. To complete this someone will have to go into each system separately to make changes while an RPA could do this instead with speed.

More recently, RPAs have been coupled with chat bots and components of AI. This means when the chat bot finishes the RPA takes over and completes the process, making it seamless. This type of RPA is called Intelligent Automation (IA).

Robots are pieces of software or as many people think of them, physical pieces of hardware typically programmable to automate a task and then complete It repetitively. Most robots have no intelligence, they can’t make judgements and just complete instructions with 100% accuracy.

Robots combined with AI is the next big thing! The better the intelligence the smarter the robot and the more applications it has. To date, we should consider the intelligence applied to robots to be limited in comparison to real human intelligence. However IA gives robots the ability to do tasks that are variable so they can adjust and correct themselves when encountering new situations.

Defining Artificial Intelligence
Artificial Intelligence (AI) is now one of the most misunderstood terms, often liberally interchanged with terms like machine learning (ML) and mixed in with talk about robots.

One of the challenges is finding an easy explanation for AI, it is a complex subject. Some time ago, I came across a conversation between Christian Reilly (@reillyusa) and Kurt Marko (@kmarko) on twitter about the misuse of the term AI. During the exchange Kurt referenced a Venn diagram he had come across that really neatly describes the relationship between AI and ML.

AI is really the outer layer of the onion with each of the other ‘AI’ terms actually more specific techniques. It clearly shows that ML is one of those techniques but on its own it is not AI.

Why do we need to be clear about these terms and what they mean?
When discussing cloud adoption with customers I always like to talk about why they are looking to take advantage of cloud and what outcomes they are looking to drive with cloud adoption. Whilst speed and agility are often the first points, I’m now discussing automation and AI, and how businesses are looking to use capabilities offered by the hyperscale vendors in these areas.

Businesses are now looking to leverage the data they have alongside new data sources and through the application of AI (and typically ML algorithms). As they are looking to: gather new insights, build new business opportunities or improve customer experiences. They look to the cloud to access the huge compute power needed for some of these solutions or to access the ready built machine learning solutions that the hyper-scalers have built in, as they are often cheaper and a quicker route to adopting AI capabilities.

We have been helping customers with automation within their infrastructures through DevOps solutions to help automate software delivery. And we are increasingly becoming involved in data solutions, migration of large data sets and streaming of new data sources. Which is enabling our customers to take advantage of AI solutions available from our partners the cloud hyper-scalers, our database experts can help find those data points and architect new data lakes to support this next generation of technology. And our Rackspace Application Services department specialises in helping customers with database architecture and database skills.