Important: The creation of a custom GPT is only available to ChatGPT Plus and Enterprise users.
What are GPTs?
GPTs are custom versions of ChatGPT created by OpenAI users. Users only have to tell the GPT builder what they want to create in simple sentences or prompts and the builder takes care of the rest. User-defined GPTs can also surf the Internet, generate images with DALLE-3 and execute code.
Confusingly, GPT also stands for “Generative Pre-trained Transformer” and refers to the family of AI models developed by OpenAI. We do not know why OpenAI has not made a clearer distinction between GPTs and user-defined GPTs. Instead, I keep hearing synonyms such as personal chatbot, personal assistant or AI assistant. It is important to understand that a custom ChatGPT can only be used within the OpenAI platform and cannot be integrated as a chatbot on the website like a classic chatbot.
Advantages of user-defined GPTs
While users with a ChatGPT Plus or Enterprise account can already customize ChatGPT responses with custom instructions, the GPT Builder goes one step further in two ways:
- We can createa user-defined number of GPTs. There is no limit to the number of GPTs we want to create, so users can easily switch between the different GPTs and call up the appropriate previously created AI wizard for the respective application. For user-defined instructions within the classic ChatGPT, however, the number is limited to one set of instructions per user.
- Upload knowledge source files. Instead of manually copying and pasting text from each knowledge source you want ChatGPT to consider when generating a response, users can upload knowledge files directly into your GPT builder. The GPT then uses this source of knowledge, among other things, to process the requests and create the answers.
- We can also share GPTs. Once we have created the GPT, it can also be shared with others. Either completely public for all OpenAI users or only with a link to your own network.
Note: As part of the OpenAI data protection controls users can exclude their account from model training. This means that OpenAI cannot access sensitive information such as chat history and knowledge files to train its models. But as with any generative AI, there is always the risk of data breaches and opportunities for people to access their knowledge sources. So be careful what data you upload to OpenAI. Peak Privacy is a data protection-safe alternative for user-defined GPTs that is particularly popular in Switzerland and Germany. More on this in my article on privacy-safe CustomGPTs.
What examples of custom GPTs are there?
Custom GPTs have become known primarily through the GPT Store from OpenAI. Every Chat GPT user can access the GPT Store via the “Explore GPTs” button in the navigation on the right-hand side and then use specific GPTs developed by others.

Here is a small selection of my GPTs
1. branding partner
This GPT creates guidelines for communication based on links and documents that you make available to it. To the GPT.
2nd AI Language Tandem
You can learn languages with this GPT. To the GPT.
3 Frida – For relatives of people with dementia
This GPT plays a role play with you so that you can learn how to communicate with people affected by dementia. To the GPT.
How do I create a user-defined GPT?
Below is a short version of how to create your own custom ChatGPT using OpenAI’s GPT Builder. There is more detail below.
- Go to chat.openai.com and log in.
- Click on “Explore GPTs” in the sidebar.
- Then click on “Create”.
- Enter your instructions in the message bar on the “Create” page. Chat with the GPT-Builder until you get the desired results. You can test the current version of your CustomGPT on the right-hand side.
- Click on “Configure” to add advanced customizations to the AI assistant. For example, you can change the name of the GPT, further refine the instructions, upload knowledge files and set up actions.
- Then click on “Create” and select how you would like to share your custom GPT.
- Then click on “Refresh”.
The implementation of user-defined GPTs in detail
The individual steps for implementation are described in detail below.
1. log in to your OpenAI account
Important: Please note that you need a ChatGPT Plus or Enterprise account to be able to use the function of creating your own GPTs.
2. create your user-defined GPT
- Click on “Explore GPTs” in the sidebar.
- Click on “Create” there.
- The GPT builder displays a split screen: In the “Build” section, you enter your instructions for creating the AI wizard; in the “Preview” section, you can interact with your GPT as you build it to more easily determine how it needs to be optimized. Enter your instructions in the message bar on the “Create” page and then press Enter or Return.
- The GPT builder will then suggest a few things based on your instructions: a name, a profile picture and standard conversation starters. You can accept the first suggestions or ask the GPT builder to change them. Even if you accept the initial suggestions, you can always change them later.
- The GPT Builder prompts you to enter more specific instructions to optimize the behavior of the GPT. For example, if your GPT generates long answers, you can specify in the builder that the answers should remain short.
- Continue refining the prompts until your GPT returns the type of responses you want your final AI assistant to generate.
3. configure your user-defined GPT
Once you have set up the basics, you can further customize the GPT with advanced settings.
- Click on Configure
- If you wish, you can change the name and description of your user-defined GPT as usual. This is how you change the other advanced settings:
- Profile picture. Click on the profile picture. Here you can upload your own photo or use DALL-E to automatically generate a new one. If you want to specify what type of image DALL-E should create, click on Create and enter your instructions.
- Instructions. Update the instructions generated by the GPT builder or enter additional instructions or guidelines on how your custom GPT should or should not behave.
- Conversation starter. Click on X next to a prompt to remove it. Or enter a new prompt in an empty “Conversation starter” field.
- Knowledge. Would you like the GPT to rely on your company’s style guide when creating their response? Or maybe you want it to search PDFs of customer personas to give it additional context. Click on “Upload files” and add all relevant files to which your GPT should refer.
- Functions. By default, the user-defined GPT can surf the Internet and create AI-generated images. If you want it to also execute code or analyze data, click on Code interpreter and Data analysis.
- Actions. If you want your custom GPT to retrieve external information or perform actions outside of the ChatGPT platform, click on Create new action. Here you can specify how the custom GPT should use a third-party API.

4. save your user-defined GPT
- When you are happy with your custom GPT, click Create. (If you have already created the custom GPT and want to publish changes to it, click on Update).
- Choose how you want to share your custom GPT: by invitation only, to anyone with the GPT link or publicly in the GPT store.
- Click on Refresh.

ChatGPT is displayed in the side panel of the ChatGPT home page along with any custom GPTs you have created. Click on the GPT you want to use and interact with it as usual.
Note: User-defined GPTs can also make errors or not behave as configured. Therefore, the response of your user-defined GPT should also be checked critically here.
5. share your custom GPT
- Click on the GPT you want to share in the sidebar.
- Click on the name of your user-defined GPT in the top left of the main conversation area.
- Click on “Copy link”.
- Paste the link as usual and share it.

How to edit your custom GPT
While it may seem intuitive to click on the pencil icon next to your custom GPT in the side panel of ChatGPT to edit the GPT, clicking on it will only lead to disappointment. This is how you edit the user-defined GPT:
- Click on the GPT you want to change in the sidebar.
- Click on the name of the user-defined GPT in the top left of the main conversation area.
- Click on Edit GPT.
- Updates the GPT settings.
- Click on Refresh.
My offers for the implementation of CustomGPTs and AI assistants

Jump start
For individuals
In a one-to-one coaching session (video call), I will give you the most important tips and tricks for implementing a CustomGPT. You will benefit from my experience and, with a little patience and skill, you can then create your own CustomGPT or AI assistant.
Joint implementation
For start-ups, NGOs and SMEs
We will implement the CustomGPT together and you will then be able to make adjustments yourself. As a rule, implementation is a process consisting of various coordination calls and joint and separate implementation sessions. Depending on the team, the AI assistant will be ready for use in 1-2 weeks.


Jump start
For individuals
In a one-to-one coaching session (video call), I will give you the most important tips and tricks for implementing a CustomGPT. You will benefit from my experience and, with a little patience and skill, you can then create your own CustomGPT or AI assistant.
This article is also available as a podcast episode
Attention! The podcast was created entirely by my AI-Assistant based on my contribution – no guarantee for incorrect content.