Solo Episode: Moving from ChatGPT to Claude, My AI Setup Explained

In this solo episode, I’m taking you inside my real-time transfer from ChatGPT to Claude. I’m breaking down why Claude is having such a major moment, what’s changing in the AI platform landscape, and how I’m thinking about rebuilding my own AI setup with more intention.

I also walk through the basics of what all these new AI workspace terms actually mean, including Claude Projects, Skills, Connectors, Plugins, Claude Code, Cowork, slash commands, and tasks.

This episode is not about saying one tool is good and the other is bad. It’s about understanding the bigger evolution happening across AI platforms. I also share my current setup approach, including how I’m cleaning up ChatGPT memory and projects, deciding what context matters, creating new Claude Projects, building brand voice guidelines, adding Skills slowly, testing simple Tasks, and avoiding the temptation to overbuild too early.

This episode is for founders, creators, marketers, podcasters, and AI-curious professionals whose ChatGPT or Claude workspace is starting to feel messy, and who are realizing the real unlock is not just better prompts, but better setup.

Helpful resources mentioned:

Anthropic Academy:
 https://anthropic.skilljar.com

Claude prompt library and use cases:
 https://claude.com/resources/use-cases

The Complete Guide to Building Skills for Claude:
 https://www.anthropic.com

Search tip:
 Google “awesome Claude skills GitHub” to explore free libraries of pre-built Claude skills.

Ready to cut through the AI overwhelm?

Explore all my resources in one place → https://stan.store/BRXSTUDIO
Free AI Guide • 45-Minute AI Jumpstart • AI Community • How I AI Companion GPT

And if you’re a brand, event, or platform interested in sponsoring season two or collaborating around conference coverage, moderation, or thoughtful storytelling at the intersection of business, technology, and culture, I’d love to connect.

More About Brooke:

Website: brookex.com

LinkedIn: Brooke Gramer

More About the Podcast:

Instagram: howiai.podcast

Website: howiaipodcast.com

"How I AI" is a concept and podcast series created and produced by Brooke Gramer of EmpowerFlow Strategies LLC. All rights reserved.

Transcript

Brooke 0:00

​Welcome to How I AI the podcast featuring real people, real stories, and real AI in action. I'm Brooke Gramer, your host and guide on this journey into the real world impact of artificial intelligence. For over 15 years, I've worked in creative marketing, events and business strategy, wearing all the hats. I know the struggle of trying to scale and manage all things without burning out, but here's the game changer, AI. This isn't just a podcast. How I AI is a community, a space where curious minds like you come together, share ideas, because AI isn't just a trend, it's a shift, and the sooner we embrace it, the more freedom, creativity, and opportunities will unlock. Hello and welcome back. Today's episode started as me wanting to talk about my switchover from ChatGPT to Claude. For those of you who don't know Anthropic, the company behind Claude, is having a real moment. After OpenAI's Pentagon partnership drew a bit of a backlash, Anthropic gained a reputational edge as the more ethically cautious option for LLMs. So many of my friends made the switch over, which proves that choosing an AI platform is increasingly an act of signaling what you trust, who you think should hold power, and what kind of future you're endorsing. According to Ramp's March twenty twenty-six AI Index, Anthropic is now winning about seventy percent of head-to-head matchups against OpenAI among businesses buying AI services for the first time. Nearly one in four businesses on Ramp were paying for Anthropic in February, up from one in twenty-five a year earlier. And that's a huge bump. On the investor side, Bloomberg recently reported that Anthropic had attracted funding offers around an eight hundred billion dollar valuation. So the more I started researching and bookmarking these how-to data transfers and following what's happening along online, the more I realized that this is not just a personal workflow change. There's a much bigger shift happening. Claude is becoming a serious enterprise and power user platform, and the reason I think users are drawn to it is because it's easy to use, manage, and set up. So for today's episode, I'm taking you inside my own AI workflow in real time, ahead of my data transfer from ChatGPT into Claude. For a bit of my backstory, I started using both ChatGPT and Claude in July of twenty twenty-three. I was on the free version to start and by September, I had become a paying user for both platforms. And at some point, without some big dramatic decision, I became just a ChatGPT person. It became the place where everything lived: my business ideas, my podcast strategy, random questions, guest prep, content, my branding ideas, half-baked offers, my dream journals, you name it. And because I've been using ChatGPT for years now, it's become incredibly useful. But lately, it's also become kind of like a junk drawer. There's a lot of really good things in there. There are valuable things in there, but also there's old projects, outdated ideas, messy threads, random experiments, half-built concepts, and context that does not fully represent where I am now in my business. A little side note, this is why managing your memory is so important to do within ChatGPT and any AI workspace that you're using. More on that in a bit. Here we are, May twenty twenty-six, I've been way too busy to reorganize my ChatGPT process and better systematize everything, especially because I do this for others with my consulting business, so doing it for myself sometimes just falls to the wayside. I'm giving myself more space between wrapping up the season one of the podcast and launching season two in the coming weeks. Now I'm finally jumping on the Claude bandwagon. Mostly because I think it's important to stay non-committal in the era of AI, and also because I want to be able to advise you all better. And truthfully, Claude feels like a new house I can organize intentionally from the very beginning. So I've been watching this mass movement from ChatGPT to Claude since around March. People are sharing setup guides, talking about projects, co-work, skills, connectors, operating systems. You know, a year ago, asking ChatGPT or Claude a question looked pretty much the same, although I can say that I've always felt that Claude writes better in terms of marketing and sales copywriting. And up until recently, you might have been using your LLMs as more of like a vending machine. You put something in and you get something back. That was the whole experience, but in the past handful of months, that has been changing. Both Claude and ChatGPT are becoming customizable workspaces. You can connect tools, you can create projects, you can add skills, you can build workflows. You give AI rules, memory, your brand voice, you upload files, and really build out the systems and context. This is the next level of AI, and the gap between a person who uses AI versus a person who is genuinely faster because of AI is shifting. It's less about what you type in and what you get back, and more about how you set up the systems around it. We've seen all these companies race to get the smartest model. 2025 was a lot of hype news cycles between Manus and DeepSeek. We're not even touching on Perplexity Labs or Google Gems. But this year, the companies that are winning are the ones that are building the best systems around the model. And the phrase being thrown around the AI landscape to term this experience is called harness. So think about the model version, that's the engine, but harness is the car that builds around the engine. Harness is what the AI can access, what tools it can use, what rules it follows, when it needs to ask for permission, what it's choosing to remember, what workflows it can run, and what happens when something goes wrong. And the biggest thing here is that these systems have to be easy, simplified, and built around intuitive, good design. And this is why I really think Claude is having such a moment. It's not just that Claude writes better, it's that Claude feels increasingly designed around the work. From CoWork to Claude Code to the newly launched Claude Design, they are pieces of a larger shift from chatbot vending machine to workspace. And did I mention how easy it is to use? But because ChatGPT has been my default for so long, it's just starting to feel like this old house I was brought up into and just kept stuffing things inside. Claude feels like this new house opportunity. And the thing that's exciting about this new house is that I get to organize it intentionally from the beginning. And this is the opportunity I'm seeing right now for myself to make the switch. Not just to switch because everybody else is, but to reorganize the way I work with AI entirely. Because if AI is becoming a configurable workspace, the question is, what kind of system am I building around myself? Is it helping me think better? Is it helping me create better? Is it helping me make better decisions, protect my energy? Is it helping me move the right things forward? So today's solo episode might just be part one of a two-part series. Today, I'm gonna walk you through how I'm actually making the move, from ChatGPT to Claude, what I'm cleaning up, what I'm bringing over, how I'm structuring my projects, and what I'm setting up first so I don't recreate the same mess in a new tool. I'll also share all my resources that I'm leaning on for this data transfer in my show notes. The practical walkthrough of cleaning up ChatGPT, exporting the useful content, building out my Claude projects, creating the markdown files, setting up my brand voice, adding in skills, choosing the right connectors, testing out my tasks, and eventually maybe one day exploring Code but let's not get ahead of ourselves. All right, let's dive into it. Let's start with the basics. I want to mention there are many parallels between ChatGPT and Claude, and so I think that this is just gonna be a really intuitive transfer for me. For example, ChatGPT has projects, custom GPTs, apps, connectors, GPTs that are already created for you in a library by other individuals, memory, and app integrations. This translates over into Claude, which also has projects. They have skills, connectors, plugins, Cowork and Claude Code. So just a simple comparison side by side, Claude Projects are just like ChatGPT projects. Both organize work into containers. Claude Skills are somewhat like custom GPTs, where the instructions or reusable playbooks are handled, but it's not exact. Uh, but it has a similar purpose with the goal to teach the AI how to do repeatable work. Next up is Claude Plugins, and These are somewhat like those custom GPTs that are in the GPT store library or those app bundles that you see within the app sections. Essentially, they package capabilities around a role or a workflow. I've already seen Claude do this in a way more branded way. Next up are Claude Connectors, and this is essentially what ChatGPT calls apps. Both connect AI to outside tools and data. Next up is Claude CoWork, and right now there's no exact GPT equivalent, but essentially, Claude CoWork is a desktop workflow layer where AI acts across files and tools. Next up is Claude Code, and the equivalent to this is OpenAI's Codex. And these are coding tools both aimed at building, editing, reviewing, and executing code or technical work on the agentic level. Another thing you might have been seeing are all these Claude slash commands, and these are essentially shortcut prompts or workflow triggers. ChatGPT has something a little similar to this, like when you create custom GPT instructions and app actions. But slash command workflows are very specific to Claude's CoWork style and particularly. Last but not least, Claude Design, which was just newly released, is very similar to what you experience in Canva. But it's inside Claude and ChatGPT connects to Canva, but in my personal use, it hasn't done a great job with it. But Claude Design is Anthropic's move into more visual work. I can see that ChatGPT has been putting a lot more efforts into their image generators. They did recently shut down Sora, which was their video generator. And you can tell they're trying to focus more during this big upheaval of so many people moving from ChatGPT to Claude. I do want to disclaim that I'm not getting paid to share any of this. I'm not associated with Claude or ChatGPT in any way. And the best way that I learn is to teach. And so ahead of this data transfer, which I'm doing tomorrow, it's going to help me to go through all of this and explain it all to you. So let's dive deep into my setup and how I'm approaching the data transfer to Claude. I'm not trying to copy and paste three years of my messy AI history and projects into the new tool. That would defeat the whole point of me trying to reorganize my information. My goal here is not to transfer over my mess, it's to organize along the way. So I'm gonna start with cleaning up the old house. I'm gonna review my ChatGPT memory, which I mentioned a bit earlier, and this is within your account in your settings. There's a section where you can go through your memory. And I've done a pretty okay job with m- managing my memory, going back and whenever I do a rebrand or a pivot or launch new products or discontinue old products within my business, I'm making sure that my memory is cleaned up and deleting those threads that ChatGPT automatically saves as high level important information. So take a quick peek in there and see what it has saved. There might be some really outdated stuff in there. Some other ways I manage my memory is whenever I'm just coming in for a one-off thing I don't want it to remember, that is when I launch incognito phase. Sometimes when I ask it something silly and I really don't want it to remember, I will just delete the whole thread. And also, when I'm having a breakthrough moment that I want it to remember, I either rename the conversation in a very specific way, whether that's in all caps or some way where I can very quickly make sure I find it and reference it within my project, or I say,"This is really important. Remember this." And then I would see ChatGPT say,"Added to memory" when I did that. So Go take a look at your memory. You might find some stuff in there that needs to be m-managed. And so this is what I'm doing ahead of the transfer over to Claude, is I'm going through and co-combing through this, trying to maybe delete some old chats, memory threads, and old projects that I, I no longer find relevant. I might even comb through my instructions for some of my custom GPTs. Some of them might need to be updated. My reasoning behind all of this is I want to see what still represents where I am now and what's outdated. Next up I'm gonna choose what matters. I'm only keeping relevant businesses, projects, and goals related to my podcast and my brand, and my workflow. So not every old idea needs to come with me. This is what I'm going to use to create these new rooms. I'm building Claude projects around my real business areas. For me, that's gonna look like BRX Studio, which is my digital storefront on Stan Store. The podcast and launching for season two and all of my goals for that. The organic content and branded content that I make for my social media and my newsletter, my social media calendar and strategy. Next, I'm going to write all the house rules. I'm adding project instructions and brand voice guidelines. This is where I'm going to tell Claude how I want it to think, respond, write, challenge me, and support the work inside each project. I've already created a lot of these for my custom GPTs and projects, so this will be an easy transfer. Once everything is updated and prioritized within ChatGPT, I'm going to have ChatGPT make markdown summaries and highlight key documents to bring over in the transfer. I think this is really important. It's not going to be everything, just the files that actually help Claude understand the work I'm prioritizing right now. After that transfer is done is when I'll start to go in and create the skills for repeatable workflows. And as a reminder, a skill is basically a recipe or playbook for how Claude should do something again and again. For example, my first skill will likely be something like a podcast episode show note writing skill. It's something I do every week and would save me a lot of time. Then I'll add connectors slowly and intentionally. I'm not connecting every app just because I can. I'll probably just start with one or two, maybe three useful ones. Definitely Gmail. I'll connect my Notion. I don't really pay for a lot of Google Drive storage. So maybe we'll just start there and see how far it gets me. Next up, I'll run some simple tests and try out some simple tasks before I create any recurring automations. For example, I'll do a trial run and say something like,"Summarize my emails today and flag anything urgent." Next, I'll add in my shortcuts, those slash commands, which essentially are just shortened prompts or shortcut buttons for workflows I know I repeat daily or weekly. You won't need a lot of slash commands unless you know what you're repeating daily or weekly already. This is why it's so important to take the time to map out your workflow and create those SOPs or standard operating protocols within your business. This really is the magic where you start to save time with AI in the very beginning. So that's it. That's how I am starting my data transfer. What I'm not doing yet is connecting every app, installing every plugin, creating ten skills right away. I'm not building out my operating systems immediately. I'm not moving every single ChatGPT conversation over... I'm holding back on a ton of slash commands before my workflows are clear and in place and working, and my test runs are working. Last but not least, I'm not canceling ChatGPT until I feel completely comfortable with this transfer, and that I'm not losing anything important. If you take away anything from this episode, I hope your action step is simple. To open the settings of whatever AI tool you use most and configure just one thing. Configure one project, connect one tool, write one instruction file, build one reusable skill or custom GPT. Don't try to overhaul everything. The real question is, am I using this AI and building a system around the way I actually work? And for me, that's why this Claude transfer is becoming so interesting, because I'm not trying to transfer the mess I'm currently in right now. This is my next real phase of using AI. It's learning how to build the system around the way I think and the way I want to work. Last but not least, I want to share some helpful resources. For example, Anthropic has a free academy. They offer courses covering things like Claude 101, APIs, Claude Code. They give you official certificates. You know, there's no longer a need to pay for two thousand dollar AI boot camps to feel like you're staying abreast of everything that's coming through in this space. And you can just go to anthropic.skilljar.com to access those. They even have something cool called AI Fluency for Educators, and this is geared towards faculty and educational leaders. They even have an introduction to agent skills. I also want to resource you with claude.com slash resources. Within there, you'll find a prompt library of use cases to create out your brand assets. So things like a newsletter writer is already set up for you, which is awesome. Another tip I recently found for resources is to simply Google awesome Claude skills on GitHub, and it'll pull up a free library of pre-built Claude skills within GitHub. Now, this is awesome. Last but not least, there's a complete guide for building skills that Claude made this PDF. I'll be sure to link these resources in the show notes so you can reference them. So let me know if you'd like a part two of this series. I can go into more deeply how I'm setting up Claude like a business co-founder. All right, everyone. I hope you enjoyed this special solo episode hope it was helpful. I love to hear back from you of what's resonating. And thank you again for tuning in. I'll see you next time. Take care. Have you just started exploring AI and feel a bit overwhelmed? Don't worry, I've got you. Jump on a quick start call with me so you can walk away with a clear and personalized plan to move forward with more confidence and ease. Join my community of AI adopters like yourself. Plus, grab my free resources, including the AI Get Started Guide. Or try my How I AI companion GPT. It pulls insights from my guest interviews along with global reports, so you can stay ahead of the curve. Follow the link in the description below to get started.