By now, almost all of you are using AI in some kind of context. It could be from finding recipes for sourdough bread to creating a sustainable hydroponic farm and anywhere in between. But there are times when you will find you are doing the same task over and over, and just not getting the right result. This can be for a series of different reasons, from using a slightly different prompt every time, to “drift” or “hallucinations” of how the LLM sees you based on interactions.

It could also be a model change, breaking much of your hard work.

All that said, there is a better solution for these often-replicated tasks where you can save the time of having to dig up your instructions every occasion you want to dial in that social media post or analyze a CMA. Make a GPT, just for that one task, and toss in your documents, photos, or metrics. Boom… a consistent and repeatable result (most of the time, this AI stuff can be weird).

What is a custom GPT?

A custom GPT is a specialized version of a large language model that is tailored to perform specific tasks by being trained on particular datasets or given detailed instructions and knowledge bases. These customized versions can handle niche use cases with greater accuracy, such as a chatbot for a company’s internal knowledge base or a marketing assistant that writes in a specific brand voice. Creating them can often be done through a no-code interface by providing specific instructions, documents, and knowledge files.

Why a custom GPT vs just talking to an LLM like ChatGPT or Gemini?

Specialized knowledge: They are trained or fine-tuned on specific data, allowing them to provide more relevant and accurate responses for a particular topic or purpose.

No-code creation: Many platforms allow users to create custom GPTs through a conversational interface without needing to write code.

Configurable instructions: You can give the GPT specific instructions, a role, and goals to guide its behavior and ensure it meets your needs.

Knowledge retrieval: They can be equipped with external knowledge bases or files, allowing them to access and use information beyond their initial training data.

Integration: Some custom GPTs can be connected to other applications and services through APIs, enabling them to perform actions like scheduling or accessing specific data.

Persistent context: They can remember their purpose and context, so you don’t have to re-explain the task each time you interact with them.

Where can I create Custom GPTs then?

ChatGPT’s pro version includes the ability to create and edit “Custom GPTs,” though some functions may differ from others.

Gemini pro, as provided through your KW G Suite account, has the capability as well, though they call these “Gems”.

The process with both platforms is similar in creation, though I found ChatGPTs version easier to follow during initial building, while Gemini you have to know more about what you are after since it does not have a conversational interface.

Finished GPTs can be private, or shared with others, based on similar sharing rules to what you would find in something like Google Drive.

While there are likely a hundred million ways you can use a GPT, on November 19th for our regional Social Media class, Hashtags to Home Sales, we’ll be showing you how to use AI tools such as Gemini to create a custom GPT which not only saves you time posting, but helps you dial in on your client avatar while remaining consistent with your personal brand. This can help you post more consistently, and with a high level of quality than otherwise, hitting all the key points in each post for engagement and responses.

Be sure to register and check it out: Hashtags to Home Sales 11/19

See you soon!

~Patrick