…Hello everyone and welcome back to the Kanopy. I'm your host and founder of Clemelopy, Jen Shannon. If you're listening, I just want to say thank you so much for being here. It means the world to me. Today, we are diving into part two of The Story, The Layoff, and the Future.
If you didn't listen to part one, that's totally fine. You can always go back and listen or you can just pick up right here. Before we jump into part two, let's do a quick recap of part…
one. In part one, I took you all the way back to two thousand and eight where I made the decision to start my first business from my couch in Chicago, building a photography career that eventually became a full studio, and then making the difficult decision to walk away when success stopped feeling like success. We talked about closing the studio less than a month before my son was born, navigating my son's unexpected heart defect, experimenting with art markets and side hustles and what it felt like to create from a place of financial pressure versus creative freedom after landing what I considered my dream job. And then came the post shower moment idea, because that's where all good ideas come from. Right?
Most of my good ideas actually come from when I am laying my head down on the pillow to go to sleep. But this time, it was a post shower moment That became Comellope. I formed the LLC, planning for a nights and weekends side hustle, and then on Halloween, I got laid off. We talked about all the shame and confusion and frustration that came with it. And that is where we left off.
Now in part two, we shift from the emotional origin story to the strategy. What did I actually do next? How did I forge the road ahead? What worked? What didn't?
And how did Carmelo be moved from idea to execution. Let's get into it…
This episode is brought to you by Clemelopy be. If you want to start optimizing your website for AI search, I put together a free resource called the twenty twenty six GEO Playbook. It walks you through how to make your content AI ready step by step and in plain language. Download it free at Clemelopy dot com forward slash playbook…
So as I said in part one, there I was…
Yep. There I was. Now an unemployed freeloader with a new business side hustle that suddenly seemed so bright, vibrant, and exciting. But the immense pressure I felt to get it off the ground was now at the forefront of my mind and feeling like a race against the clock. I needed to have it all figured out in a very short period of time because now my regular paycheck was gone.
We had severance and savings in the bank, and I didn't wanna touch any of it. The first thing I did was create a basic Squarespace website and file my trademark application. If you don't know, in order to file a trademark application for a mark that you want to use right away, as in not an application for intent to use, you have to show that you are using that mark in business already. So the basic website accomplished that for me. Now let me give you a disclaimer.
I am not a developer. I'm not a software engineer. I'm not fluent in most coding languages. All of that. But I've always had the mindset that if someone can be taught to do something, then I can learn to do it.
And that's how I've always done things, from learning to play piano when I was young and again a few years ago, to teaching myself guitar in seventh grade and playing really well for fifteen years, to being a cake decorator at Publix. I even taught myself how to woodwork and turn things on my lathe. The point is if I wanted to do something, I was going to find a way, and I was going to use whatever resources I had at my disposal. And I know a lot of people will say that in the famous words of t swizzle, I didn't do it legitly. But I'm sorry.
Who are you? The learning police? The learn by doing is my philosophy. We learn by making mistakes. We learn by cultivating information and putting it into practice, and a university degree isn't necessary for that in many cases.
So don't listen to the people who say you didn't do it legitly. The point is you're doing it. You're learning, you're failing, and you're succeeding, and the most successful people leverage the tools they can access. So for me, that started with my experience of seventeen plus years and ChatGPT. The next thing I did was use ChatGPT to help me find the best tools to use to develop what I was trying to develop.
I had this idea in my mind and I told it to ChatGPT and asked it for solutions that I could use to bring it to life. The first solution I landed on was Softr. As a brand new founder, I was naive in a lot of ways, but one thing I was not naive in was knowing that this would be a huge learning process with a big learning curve. I knew I'd make mistakes and fail at a lot of things. And no matter what, I would course correct and find my path forward.
So when I started digging into softer and learning the ins and outs of building with it, I very quickly realized that it wasn't going to work for what I was trying to build. I couldn't deliver a truly native Clemelopy experience. The portal's behavior was highly dependent on softer quirks. For example, it didn't handle Boolean or text values cleanly, and instead, it needed helper files that stated true and false to make the filters behave. In addition, I was using Airtable as my database, and I didn't really know much about Airtable, but I knew it was a database and figured it since it had a free option, I could try and use it.
But I quickly realized something extremely important. What I was building wasn't secure. And in order to make it secure, I couldn't use just Softr. Anything involving AI or API keys would not safely live in Softr. I would need a backend layer to keep secrets server side, same with the environmental variables.
And in addition, data access control depended on page by page filtering rules. So, for example, a, quote, unquote, secure portal approach in Softr was basically programming it to only show records where a Airtable email equals the logged in user's email. But the problem with this is that if any one thing is misconfigured, you can accidentally expose other customer information and records to the wrong user. And it didn't enforce a database layer with row level security, which is highly important. I learned I'd have to go up to the two hundred dollar per month plan, which wasn't doable for someone who just got laid off.
On top of that, when I eventually tried to pull the trigger on the two hundred per month plan, it turned out to still not be able to do what I needed because it still didn't allow me to embed an AI experience with custom back end control or enforce that data layer. So luckily, I was able to get out of that because I was on a free trial. So I canceled it and softer was out. The next solution I tried was WeWeb. My husband can back me up when I say this, but the two weeks I tried to make it work with WeWeb were probably the most inferior infuriating…
two weeks of my professional life. At first, it seemed like it was going to be a powerful option. It connected to APIs. It supported custom databases, which I had already decided to go with Supabase after leaving Softr…because it allowed custom logic. But as I began building, I realized not only was what I was building more complicated than what WeWeb could handle, but the debugging was the worst.
Logic…for how you build it or how what I was trying to build lived in visual workflows, which meant you literally needed to build a workflow for every single thing. And having more than one workflow for the same object could cause a domino effect of breaks. And if you tried to use their AI helper to help, it would mess it up even worse. And errors were extremely difficult to trace because, you know, in code, you can trace it, but on a front end link WeWeb, it's not as easy to find. So there was this, um, ambiguity, if you will, on the workflow level.
So in addition, even though WeWeb integrates with Supabase, much of the enforcement still depended on front end logic. So, like, visibility and conditional rendering lived in the UI, and it was easy to accidentally expose endpoints and API routes, and it still had to be secured elsewhere. And nothing was consistent. The database auth plugin wouldn't authorize and even their own chat bot said it was a known issue they were working on. But the article it showed me was from twenty twenty three.
This was November twenty twenty five. So the two years it's been sitting out there as a known issue and it hasn't been fixed. Now I tend to be a pretty laid back person, but this was the angriest two weeks I've ever had in my professional life. I think…
I I don't know. I think I've had…a lot of, you know, frustrating experiences as an adult, but this one was one where I would fix one thing and somehow it would break three other things. Then I tried using, like I said, it's built in AI to help move it along or fix whatever was, I couldn't even figure out what was wrong, but I knew something was wrong and it would just break everything else. So in a desperate attempt to salvage my sanity, I went to Claude…and told it what I was trying to do and asked for its recommendations. Now before I go any further, I want to say that I started with ChatGPT…But unless you're working with ChatGPT and Claude and Perplexity on a normal regular basis, you don't really learn the ins and outs of each of these generative engines and what they're good at and what they're really not good at.
And ChatGPT was not at all helping me to get where I wanted to go. And I had heard that Claude was really good with coding. So I thought, okay. Well, let me go over to Claude and see how Claude can help me. So that's where I pick up from.
It came back with this setup that I currently have now, which is using Next. Js. Now Next. Js is not a drag and drop editor or a builder. It's actually a real development framework.
And while, yes, you can use tools like Figma to design things first, it was yet another thing to have to figure out and learn. So I decided to just leave Figma to the side. So I did what I always do, though. I dove into learning more about the coding I needed in order to get this thing off the ground. I know HTML.
I know CSS. I know enough to get by with SQL and some JavaScript, but I wasn't good with JavaScript, certainly not TypeScript because that was a foreign concept to me. And working with Next. Js and React were also completely a foreign concept to me. But my friends, this is what ended up being the winning combination.
I wanna take a moment to elaborate a little bit more on Claude. As I mentioned, uh
, just a few seconds ago, that each of the AI engines have their own strengths. And so I wanna take a moment to talk about Claude. For those of you who don't know, Claude is an AI engine created by a company called Anthropic. Claude is known for its superior coding abilities compared to ChatGPT and perplexity, and…I thought I'll give it a try fine. So I remember the first conversation I described exactly what I wanted to build, all of my frustrations with ChatGPT and Softr and WeWeb, and…that I already had an established brand.
I told it my brand colors were orange and teal with a touch of green and gave it the hex codes. I gave Claude, um, some screenshots of UI examples that I liked, which included glassmorphic cards and frosted glass, and it gave me a mock up that was beyond anything I had imagined. And while it isn't fully what you see today, it was a sweet jumping off point to get to exactly what I wanted and to what it looks like today. And we are not done because we are always looking to improve and add to and make this a wonderful experience for our users. But I'll save that for another time.
This is the part where I refer back to what I said about people who say you didn't do this legitimately. Again, it's the scrappy people who use the tools at their disposal who make things happen. We don't just follow the rules just for appearances sake. We forge our own way. And if you think of it, it's kind of like MacGyver style.
Like, he gets put in a sticky situation, and he finds a creative way to get out. And, really, we have to forge our own way in order to create the vision we have in our minds if we're not somebody who's knowledgeable in certain areas and certainly don't have the income…
to hire people to do that. And so Claude was my tool, and it still is. We've had hundreds of conversations at this point. We've built a login system, a dashboard, a content management system, a billing system, a support chatbot named Penelope after My Begonia, um, a schema markup builder, my ticketing system, page audit…audits, and an entire framework that educates and guides users through optimizing and building their content ecosystem…
And all of this is through my ideas. These weren't ideas Claude or AI tools gave me. These were my ideas trained with my intellectual property and using tools to help me build the vision…in an area that I'm not as strong. And so Claude helped me every step of the way from building the authentication system system to incorporating Google sign in…all the way to ensuring each table had the correct row level security in Supabase and creating the middleware that checks if someone's subscription is active before letting them into the app, and even finding the perfect resource for building workers, which I proudly use CloudFlare for, and creating gorgeous PDF outputs from our first product, which was the linking strategy map. So Claude has been such a huge help for me, and I wouldn't have been able to do this without being able to learn what I learned by using Claude.
Now since getting laid off on Halloween afternoon, every single day has been spent working…every waking hour that I wasn't running kids around or making dinner. Some days were sixteen hour days. Most days were at least twelve hour days, and there was no time off. There were no weekends, evenings, or holidays. There was only me to move things forward.
There were hard days. I remember spending an entire afternoon trying to figure out why the PDF output was cutting off content on certain pages. And then Claude and I went back and forth debugging, testing, iterating, all of that stuff. We tried using a system called cloud convert first, then migrated to Cloudflare's browser rendering, which actually was cheaper and gave us better control. And then there were breakthrough days, like the day Stripe billing integration worked for the first time, the day I saw my first real user submit a project, the day the confetti animation fired after someone generated their linking strategy, And I literally cheered out loud in my office because it was amazing.
It was like finally getting to see it all come together. But what I want you to take away from this is that I'm not a developer. I'm still not a developer, but I learned that you don't have to be. You just have to be willing. And by willing, I mean, willing to learn, willing to fail, willing to ask for help, willing to use the tools at your disposal.
And in twenty twenty five, twenty twenty six, and beyond, sometimes asking for help can mean partnering with AI tools like Claude to fill in the gaps in your knowledge. And no. As of now, I am not an affiliate for Claude. I don't even know if they have an affiliate program, but this is what I have used…
to build this company. And the thing is, Claude remembers our context. So not every detail is something that it remembers right off the bat, but the important stuff. So when I say let's work on glassmorphic styles, Claude knows I mean the specific hex codes, the blur values, and the border opacity we settled on after twenty billion inter iterations of it. And when I say let's update the orchard builder, Claude knows that it used to be called the linking strategy and why we changed it.
And when I say I'm writing a podcast episode about how Camilla became to be, Claude pulls up our history and helps me remember the journey. So it's just been really incredible to have something at my disposal that is relatively affordable and that I can get results with about for now. So just before Christmas, I decided I had come far enough along to announce a beta launch date. I chose January thirteenth because my oldest daughter's birthday is the twelfth. And if you know anything about me or Clemelopy, we are huge Swifties.
And yes, Clemelopy, the inanimate business is a Swiftie. Her favorite album is The Life of a Showgirl. I mean, can't you see it in her colors? Come on. Anyway, I figured if Taylor's lucky number was thirteen, then maybe it could be Clemelopy's too.
So I decided on January thirteenth, it was really a matter of reverse engineering what I needed to do in order to be ready to go live on that date. And the first product that we had that I was working on was the linking strategy map, the to do system, and the workspace itself, kind of all in one with the promise of a GEO framework. And…
as you can imagine, as you begin developing ideas and systems, sometimes they kinda take on a life of their own, and it's different than anything you imagined. And so Clemelopy was no different. While the linking strategy was only part of the equation, I wanted a solution that was all encompassing, but I knew that would take more time. So I had to decide to start somewhere, and that's what I did. So in that time, I finished developing what was going to be launched on January thirteenth, and I apologize for my voice getting croaky.
It happens this time of year with all the pollen. Good times. But, anyway so…in order to…get that going, again, I finished developing what was going to be launched on the thirteenth, and then I created a mailing list, a beta tester program, which eventually evolved into being a public beta. I purchased insurance. I started posting to social media, reaching out to local business owners for case studies.
Um, I created a support ticketing system and tied up all loose ends with my alpha testers, did all my OAuth application stuff, connected my Stripe account, set up a business checking account, and created a single member LLC operating agreement so I'd have good LLC hygiene, you know, all the stuff that goes into a business. And and the night before launch, I stayed up into the next day of the launch…
to get everything perfect. So when I went to publish, it was like two AM on the thirteenth. And I think I was supposed to be launching at, like, ten AM. And when I went to hit publish, something went wrong. And that something undid everything I had just spent the entire night working on.
And I ended up staying up until almost six in the morning trying to fix it all to have it ready to go live at, I think, it was ten AM, but I can't remember exactly. I just remember my husband took the kids to school for me that day, and I went to bed and…got up at eleven and went straight back to work. But one of the funny things is that because Claude remembers our history, I went back to Claude and I asked it to pull up that memory. And so here's what it says. It says, oh, Jen.
I found it. Yes. I remember. It was around three fifteen AM on January thirteenth. You had tried to use chat GPT while I was down, and it created a complete mess with merge conflict markers scattered throughout your files.
Your work was tangled up in those markers, and when you tried to push it to GitHub, everything got confused. I remember you saying I'm so beside myself right now. I can't believe it. It's three fifteen AM. I thought I was going to go to bed, and now I have to spend hours on this again in order to make my deadline.
And I am so tired. And I told you the truth. Your work wasn't gone. It was just tangled up. I suggested you go to bed, that the old version of the site was still live and your customers wouldn't know the difference for a few hours.
But you found a copy of the folder with the mess in it, and we worked through cleaning up the merge con conflicts file by file. There was confusion about which GitHub repository was connected to which Vercel project. The package dot JSON still said, Clemelopy website instead of Clemelopy beta. The git remotes were pointing to different repos files that said they were pushed weren't actually making it to GitHub. And you pushed through until almost six AM.
Your husband took the kids to school that morning. You slept until eleven and went straight back to work. I'm sorry that night was so brutal, but you made it through. Isn't that funny? I love how Claude wrote that for me.
So that's what it remembers about that night. But the amount of hours that goes into being a founder of a startup is sometimes unreal. At one point I could tell my husband was getting tired of me only emerging from my office to come eat dinner, which quite frankly, he was already annoyed by because I'd say I'm coming. Just let me finish this one last thing because I really did need to get to a good stopping point. And then fifteen minutes would go by and everyone at the table gave up waiting for me.
Then I'd eat, give my lame excuse as to why it took me so long to get to the table, and then go back into my office and close the door. I told my husband, this isn't going to be forever. I just need to get past launch. And I think about that version of me, the one sitting in her office at three fifteen in the morning, exhausted, frustrated, definitely crying, and typing the words, I'm so tired, to an AI engine because there was no one else awake to say it to. There may have been a few f bombs in there as well.
I'm not gonna lie. And I want to tell her, you're going to make it. Not perfectly, not gracefully, but you're going to make it. And by make it, I mean create it. And that's the thing people don't realize going into starting to build a business.
I've had many of these nights since I started my first business in two thousand eight, and I knew that going into this. I knew I wouldn't be immune to it, but this was by far the most difficult emotionally…
because it felt like the two and a half months I spent building up to that point in an instant was ruined. It's not just strategy. It's not just systems. It's you alone at hours. No one else in your household is awake to witness, choosing to keep going anyway because you have a goal.
You have determination and grit. And because you made a promise to yourself that you refuse to go back on because you believe in keeping your word and you're a man or woman of integrity. So if that's you right now, if you're in the middle of your own three AM moment, just know that you're not alone even when it feels like you might be. But…in the end of the day, if you're building something, just know that you will hit these obstacles. You will run into walls that you have to then maneuver around somehow, but don't give up.
Because at the end, you will be able to look back and see…what built you into the person that you become and the business owner that you become…
from that journey. So that is the history leading up to the beta launch on January thirteenth. And so when we talk about the future, we talk a lot about what's going to happen, what we want to happen, and so on and so forth. But what I've learned through this journey is that…your initial plans, your initial visions can sometimes turn into something so much bigger and so much better…than you ever thought possible. And sometimes you don't have what you need in order to make those things happen.
But, again, if you can use the tools at your discretion or at your disposal that you have access to in order to build something amazing, then don't let the idea of, oh, well you didn't go to school for that or something along those lines…stop you from doing it. Because this is your life. And we all do things differently. We all walk different paths. So I hope that encourages you.
Now, as for the future of Coomelope, we are still very, very young, and I can't wait to see where we go. But I am so excited to continue building this, even with all of the trip ups that I have had along the way. In the next segment of Diary of a Founder, we'll talk all things beta launch. But until then, keep growing forward
…