4 min read

Automation and AI - What's the difference?

Picture of a robot loading washing into a washing machine.
What is the difference between AI and Automation?

Automation and AI are two distinct things but they are complementary, as AI turbocharges the potential for automation in organisations and businesses.

What is Automation?

In the simplest sense, Automation is using a machine or device to replace something that a person would normally do in several steps involving thinking and conscious actions.

This doesn't need to be a computer. Arguably a washing machine is a good example of something that provided an automation solution. Rather than someone (probably a woman) having to spend hours manually hand-washing clothes, they can load the clothes into a washing machine, press a few buttons and the machine will wash the clothes without any action required.

💡
Fundamentally, Automation is when a task goes from needing regular attention or effort from a human to when it can left alone to be performed while the human does something else.

Problems with Automation

While Automation can save people a lot of time, it has traditionally been limited in that it typically needs to have been used with only very specified and logical tasks.

For example, in the washing machine example, someone still needs to sort which clothes need to go into the washing machine and put the clothes out for drying or loaded into a Dryer. These kinds of tasks are harder to automate because they require a more human-kind of intelligence to perform. Humans have always been much better at vaguer, messier tasks than machines. At least until recently...

What is AI (Artificial Intelligence)?

This is where Artificial Intelligence (AI) comes in. The best way to think of AI is as a type of computer program that simulates how human intelligence works. Despite what you might think, generally humans do not think in the same way as computers do, which allows us to deal with problems that machines can't deal with, but often in a messy, unreliable way, often based on intuition and guessing.

AI models work in a similar way to how human intelligence works by being trained on huge quantities of examples and looking for patterns and learning how to find the "best" solution to a particular question or input.

It's important to mention that AI models can have many uses apart from being used for Automation, including text, image and video generation and editing.

Why is AI so good for Automation?

Artificial Intelligence can fill that gap in Automation where certain tasks needed to be done by a human because they weren't tasks that could be done without human-like intelligence.

So in the washing machine example, potentially robots could use artificial intelligence trained on picking up dirty washing from around a house and putting it into a basket and then putting it into a washing machine.

The same is true with a lot of computer software. Automation tools have existed for several years that let competent non-programmers create automated flows to do things like put names from emails into a spreadsheet and automate booking of appointments from a website, but they have been fairly limited because of the need in most processes for something requiring an element of human-like intelligence.

Now with software like n8n that supports including AI nodes with-in their existing processes:

In this example, on receiving an e-mail on Gmail we can send the content of that e-mail to an AI Agent node (using ChatGPT) and ask it to summarise the e-mail and tell us whether it is something we need to reply to by sending us an e-mail (you wouldn't do it like this in practise...)

You give the AI a "prompt" in a text form to tell it what to:

Summarise the following message and include whether you think the email needs to be responded to:
{{ $json.text }}

The {{$json.text}} is replaced with the text in the e-mail. Generally AI accepts instructions in a text-based form called a "prompt", though increasingly, it can take other inputs as well (for example images or even videos).

I've heard about AI agents what are they?

Despite being called an AI Agent node in the n8n flow, what most people are referring to as AI Agents are processes where AI handles the entire workflow itself outside a tool like n8n.

If you had a washing robot you would just ask it to "do the washing" and hopefully it would go and do the washing. The issue I have with AI agents, at least at the moment, is that very often they can't manage to do tasks consistently without occasionally doing something strange or unexpected.

In conversations I've had with people using products like n8n for businesses, most of them want to keep the AI element of their flows to the minimum, at least for the moment.

Conclusion

So Automation is the process of replacing tasks that need human input and attention with something a machine can do. AI helps Automation tools do things that previously required a human-like intelligence, allowing much more automation of tasks than previously was the case.