-
What I tried: Using both ChatGPT Projects and Custom GPTs for different types of work.
-
What I learned: Projects are great for exploration and ongoing conversations, while Custom GPTs are designed for repeatable roles and specific jobs.
-
How you can apply it: Start with a Project when you’re figuring something out, then build a Custom GPT once you know the exact role you want AI to play.
If you have a ChatGPT account, you may have noticed two features that sound similar but serve very different purposes: Projects and Custom GPTs.
At first glance, the difference can feel subtle. But once you understand how each one works, you start to see how powerful they can be.
Let’s start with Projects.
A Project is basically a dedicated workspace inside ChatGPT. You can upload files, have multiple conversations, and maintain context around a particular topic or problem you’re working through.
I use project for all kinds of things, from health to brainstorming content. Recently, I’ve been using a Project for something personal. I injured my calf muscle right where the Achilles attaches. So I created a Project where I uploaded my physical therapy notes, tracked my progress, and recorded what exercises I was doing each day. I could tell it things like, “I ran two miles today,” or “This exercise felt tight,” and keep the entire recovery conversation in one place.
It’s a great tool for exploration, brainstorming, and ongoing context.
Now let’s talk about Custom GPTs.
A Custom GPT is different. Instead of an open workspace, you’re building an AI for a specific role.
For example, you might create a sales training GPT for your team. You could upload your product manuals, sales scripts, philosophy, and role-play scenarios. Once it’s built, your team can interact with it anytime to practice sales conversations or review product knowledge.
You could also build Custom GPTs for things like:
-
Writing blog content in your brand voice
-
SEO page analysis
-
Strategic brainstorming
-
Customer support guidance
- HR & Employee Manuals
In other words, a Custom GPT is less like a notebook and more like hiring a specialist.
Often the best approach is to start with a Project. Use it to explore ideas, upload resources, and experiment with conversations. Once you understand what role you really want AI to play, you can then build a Custom GPT designed for that exact purpose.
Think of Projects as your workshop, and Custom GPTs as the tools you build once you know what you need.
This matters because the real power of AI isn’t just asking random questions. It’s creating structured environments where AI can support your thinking and your team’s work in a consistent way.