What Took Me Years to Fail At, I Finally Built in 10 Hours With AI
Not an AI expert? Neither am I. Here’s how I still built something real in record time.
From posts to possibilities
Over the past 18 months, I’ve built a solid community on LinkedIn: more than 21,000 people, most of them product leaders or in related fields.
What started as me simply sharing my thoughts turned into something bigger. I discovered a real passion for simplifying complex product topics into visuals — and it resonated.
More and more people reached out to me:
telling me that reading my posts felt like a relief,
saying they could heavily relate to the struggles I described,
asking if they could reshare my illustrations,
and even wondering if I’d ever write a book or coach them.
For a year, those questions kept circling in my head.
But here’s the truth:
I’m not a writer.
I don’t have the time (or appetite) to build a side business, write a book, or set up elaborate marketing funnels.
Still — when enough people keep asking for more, you start to wonder: is there an opportunity here to offer deeper content, maybe even monetize it?
That’s when I decided to give it a try. Here is my journey.
Read this only if you’re not an AI expert
Let me also start with a warning: if you’re already beyond beginner level with AI, this piece is not for you. 😂 You’ll laugh out loud at my excitement. You’ll roll your eyes at how amazed I was by things you take for granted.
But if you are a beginner, like I am: I’ve been using AI daily for the past year — ChatGPT, Gemini, and smaller integrations inside the tools I work with. It’s become part of my workflow. But I had never built anything with it. Not end-to-end. Not something that felt like a real product.
That changed with this project. In less than 10 hours, I went from scattered content to a polished guide, interactive assessment, automated email flow, and landing page — all with AI as my co-pilot.
This is the story of how it finally clicked.
Disclaimer
Before I dive in, a quick note: I don’t get paid to mention any of the tools I used here. These are simply the ones I picked after testing around. They might not be the best for everyone.
My requirements were pretty clear from the start:
I didn’t want to spend more than 50 CHF.
I didn’t want to code myself.
I wanted to automate the flow end-to-end as much as possible.
And I wanted it to look acceptable enough to be proud of, even if I am not a skilled designer.
That’s how I ended up with Typeform, Gamma, and Google Apps Script. Different tools might fit your constraints better — this just happens to be what worked for me.
Step 1: Uploading my brain
I had 17 years of product leadership experience bottled up in:
300+ LinkedIn posts over the past 2 years,
over 60 illustrations
learnings from leading teams in various industries.
notes from talks with countless product people globally
All scattered. All unstructured. So I did what most of us do today: I dumped it all into ChatGPT.
Instead of staring at a blank page, I now had a sparring partner that could hold the entire mess in context.
We went through dozens of iterations:
“Cluster my posts into categories”
“Find common themes”
“Make this section more actionable.”
“Expand this into a full page.”
“These two challenges overlap — merge them.”
Piece by piece, the chaos turned into something coherent: a structured guide around the hardest problems product leaders face - at least from my perspective.
Step 2: Making it beautiful (and failing forward)
Content alone wasn’t enough. I wanted the guide to feel like something people would actually enjoy reading.
Naturally, I started with my go-to tool: Canva. I’ve been a Canva fan for years — it always made me feel like a professional designer. All my illustrations live there. It’s simple, intuitive, and I’m still on the free plan. But designing an entire guide in Canva? That was a different story. Keeping it from looking like a wall of text turned out to be way beyond my patience.
Frustrated, I switched to Google Docs. Not because I believed Google would magically make my guide beautiful — but because I needed structure. My plan was to then hand it off to a professional on Fiverr to polish the design. But quickly I realized: finding someone I trusted, who could do it well and stay within my budget, would be harder than I thought.
I went back to ChatGPT: “What tool would let me design something closer to a book or a microsite, but without hiring a designer?” That’s how I found Gamma, where created a free account, and the onboarding was seamless. Within two minutes — thanks to their AI assistant — I had something that already looked far better than anything I’d managed on my own. It still took plenty of hands-on refinement to reach the polished version I envisioned, but the energy shift was huge. My biggest hurdle — making it look good — was suddenly solved.
And just like that, the guide wasn’t just words on a page anymore. It looked polished, intentional — something I was finally proud enough to share with my husband for feedback.
Step 3: Designing the assessment
I have to admit: I’m not a marketing person. For years, I’ve followed countless influencers. I know the theory behind lead magnets, funnels, and audience targeting. But in practice? I always struggled to find the right tools — and to tie them together without sinking hours into manual work.
This project felt different. Not because it was easy — it wasn’t. It was complicated, often overwhelming. But my mindset had shifted. I’ve always believed that as a leader, you need to walk the talk. And with so much now possible through AI, I decided to give it a real chance. This time, I was determined to figure it out.
As a starting point, I thought an assessment could be a good entry into the journey. I’d built assessments before, and they had worked surprisingly well. So I set out to design one for readers to self-assess their leadership maturity.
With my LinkedIn content and the guide as a foundation — and ChatGPT as a sparring partner — I created a 15-question test. Five key dimensions. Each answer rated on a scale of 1–5.
The scoring was simple: results would fall into three categories, and each category linked back to specific chapters of the guide for deeper reflection.
Within minutes, the whole thing was live in Typeform. Clean. Simple. Ready to use.
Step 5: The automation rabbit hole 🐇
This was the part I was most afraid of. I’m a full-time working mom, and my day-to-day is demanding. But after investing so much time into crafting the guide, polishing the designs, and shaping the content, giving up was not an option.
By now, the first pieces were ready:
The guide as the core artifact.
An assessment (created with ChatGPT and built in Typeform) so readers could benchmark themselves.
What was still missing: the automation. There was no way I could send manually crafted emails to everyone who took the assessment. I’d tried something similar three years ago — and gave up after two weeks, feeling twenty years older.
So once again, I turned to ChatGPT. This time, not for content or design help, but to discuss my idea and find the right, affordable solution.
Step 5: The automation rabbit hole 🐇
At first, the idea of automation felt daunting. I could already picture myself lost in documentation, wasting evenings trying to debug code I barely understood.
But ChatGPT turned this into a different kind of battle — one I could actually win.
I explained what I wanted: “Typeform collects the answers → they land in a Google Sheet → a script checks for new entries every five minutes → and sends each person a personalized email with their results and links to the right chapters.”
Line by line, ChatGPT wrote the script. It explained what each piece did, suggested where to paste it, and adjusted whenever something broke. I tested it over and over. At least ten times. Sometimes it worked, sometimes it didn’t — but instead of giving up, I just copied the error back into ChatGPT, and it patiently rewrote the code.
Slowly, it came together. Well - it required a lot of back and forth until the eMail looked nearly as good as I wanted it to be.
And then, one evening, it finally worked: a user completed the assessment, the script recognized the new row in Google Sheets, and within seconds an email with personalized results landed in my inbox.
It felt like magic. (To me, at least. My husband just found my excitement “cute” — but then again, I did marry a developer.)
I hadn’t become a coder overnight. But with AI as my co-pilot, I’d managed to stitch together something I had failed at three years earlier — a fully automated flow that actually worked.
Step 6: Packaging the story
Once the Typeform was ready, and the eMail automation worked seamlessly I realized something was missing: context.
What is this guide actually good for?
How does it help?
Who should take the test?
Without answering those questions, people wouldn’t know why it mattered. Which meant I wasn’t done yet — I needed a landing page.
Given my earlier experience, the choice was obvious: Gamma. Within minutes, I had a clean, simple page that explained the guide, framed the assessment, and invited people in.
It didn’t just look good (personal opinion) — it gave the whole project a home.
And because I was having so much fun, I went a step further — I even built a website. That turned into a whole different story, because along the way I realized just how many amazing people I’ve already recorded podcasts with.
Epilog
By the time everything was ready, I stepped back. Gumroad Shop. (I didn’t even write about this part) Guide. Assessment. Automation. Landing page. Website.
And then it hit me: the whole thing had taken less than three days of focused work.
For a beginner like me — someone who had tried and failed at similar projects multiple times over the past years, and who honestly still considered herself a noob — that felt huge.
Not because I cut corners. But because I had:
A small enough project to finally get started.
Plenty of content to feed the AI monster.
Trust that this time it would actually work.
A bit of patience along the way.
And — maybe most importantly — a lot of fun mastering something I had failed at for years.
Why I’m not sharing this
I’m not a natural salesperson (I’m a product person — which means, yes, imposter syndrome is part of the package). So I didn’t write this post to convince you to buy my guide. But if you do want to? Well, I certainly won’t stop you. 😇
Why I’m sharing this
The process mirrored the best of product work:
Start messy → dump everything in.
Iterate fast → don’t expect perfection upfront.
Test constantly → whether content or code.
Automate overhead → free yourself for value.
Package beautifully → people engage with what feels intentional.
There’s also a big lesson in all of this for me as a product leader — but that’s probably a story worth its own post.
I wrote 20 Hard Challenges and How to Solve Them to help product leaders face the messy reality of leadership. But the way I built the guide taught me something too:
The leverage is here. The tools exist. You just need the willingness to experiment, fail forward, and keep iterating. If you’re sitting on a folder of messy notes, half-written blogs, or old slides — this is your sign. If I can do it - you can do it too).
So what is holding you back? Share your feedback here or connect with me on LinkedIn.
* This post is Co-Authored by Chat GPT - but I promised it was not a one-prompt-post. We had quite a lot of back and forth.
Great post Stephanie! It’s a good example of taking a well scoped problem so that you can execute within 3 days. Thanks for sharing!