ChatGPT vs OpenAI API: What they are and why it's important to know the difference!

ChatGPT is what most people who've dabbled with AI are most familiar with when they hear the term 'AI'.

ChatGPT is like a really smart, friendly expert that you can talk to. You ask it questions or give it tasks like writing, explaining things, solving problems, or even just chatting, and it replies in clear, human-like language. It’s trained on a massive amount of information, so it can help with all kinds of topics, from everyday advice to technical stuff. It's like having a super helpful assistant at your fingertips. ChatGPT is the end user application that we use each day.

While ChatGPT can be massively helpful as a sounding board, research tool or writing assistant, it doesn't integrate very well with your business processes in an automated way. It still requires manual interactions and copying and pasting of data. No doubt still a massive time saver but still involving some very 'mandraulic' steps.

OpenAI is the company that makes ChatGPT. The OpenAI API, is like the engine behind ChatGPT that automators use to build their own workflows, websites, or tools that include the same smart ChatGPT technology. But instead of chatting with it directly, we incorporate it into other things (like virtual assistants, automation workflows, customer service bots, or creative tools).

What's an API?

Without getting too technical, an API (Application Programming Interface) can be thought of like this:

Imagine an API is like a waiter in a restaurant. You (the customer) look at the menu (what the API can do), place your meal order (send a request/make an API call), and the waiter (API) brings your food (the response) from the kitchen (the system behind it) without you needing to know how the food was made. The waiter provides an interface to the chefs in the kitchen.

In simple terms: an API lets two different pieces of software talk to each other and share information or services in a safe and structured way.

The difference then is that:

• ChatGPT is the finished application you use daily and interact with manually

• OpenAI API is the backend AI tool that automators and developers use to build smart workflows that are augmented with AI and you interact with it programmatically

I'm just using the ChatGPT/OpenAI example here, because its the most well understood, and because close to a billion(!) (yes, with a B!) users use it each week, but there are countless other AI end user tool/API combinations available that are similar; such as Claude/Anthropic API, Google Gemini/Gemini API and so on.

Now you understand how ChatGPT and OpenAI API differ, you can start to see how OpenAI's API (or any other AI API) can help with with our workflows.Rather than manually interacting (copy/pasting/asking/prompting) with ChatGPT, with OpenAI we can do the same kind of thing but automatically as a step in a process.

Automations augmented with AI

That's what we mean when we say "automations augmented with AI".

Automations simply put are "workflows that combine 2 or more steps to provide an outcome automatically and in a consistent repeatable way".

When adding AI to an automation workflow, we write similar prompts as we might for ChatGPT, but as code to direct the OpenAI API to do something in our workflow instead.

The OpenAI API allows us to add these AI steps into our automation workflows, passing information to the AI, without having to intervene manually - like we would have to with ChatGPT. You can augment automation workflows with AI in many different ways. Here are a few very simple examples:

1. When an email arrives matching some criteria, use AI to categorise the email in a certain way or get AI to draft a personalised response

2. An automation queries a website for some information and an AI step can take that information and organise it for you or turn it into structured information in spreadsheet or database

3. An automation can gather feedback from customers, and an AI step can analyse the feedback and determine the sentiment of the message

4. An automation that takes a report as input and an AI step summarises the report say for an executive brief

So, while these two tools important tools provide similar outcomes, it's the OpenAI API (or other AI API) that we use in our automation workflows to provide helpful and powerful augmentations that a human would otherwise have to do. It's important to understand the distinction and to know that we are AI-powered automations are truly automated.

Repeated reliable automatic AI steps in workflows without copying and pasting a single thing.