I'm a CTO at Kyan Health, a company in the mental health and coaching space. Explaining the inner workings of how we operate is not what we typically do publicly. But over the past few weekends, I completed a side project that I think illustrates something bigger than any single company: we have been genuinely augmented, and anyone willing to learn the tools can do things that were impossible a year ago.
I made an album. Nine tracks. With music videos. A website. Merch. Marketing campaigns. And I did it in weekends, over the course of a few weeks, using AI tools for nearly everything except the one thing that matters most: the message.
This is not a story about replacing artists. It's a story about what happens when a father with something to say gets access to tools that let him say it. Here's the full breakdown.
Why Music, Why Now
A year and a half ago, my son was born. It has been brutal and beautiful in equal measure. I've failed everywhere and somehow also succeeded. I work in mental health and coaching, and if there's one thing that's crystal clear to me, it's this: how you channel your emotions is a superpower.
Someone suggested I try making music as a way to process what I was feeling. At first it was just for me. I've always made music, since I was a teenager messing around with Auxy on my phone, writing little melodies that never went anywhere. But this time was different. I had something real to say.
I started thinking about the topics I wanted to explore:
- The founder-parent paradox — feeling like too much of a parent at work, and never present enough at home.
- Values and legacy — what can I share with my son? I don't want him to live my broken dreams, but I hope he'll be brave, strong, and kind.
- The immigrant experience — being a mix of cultures, feeling that the world has sometimes vilified people like us, and wanting to say: we do things right, we want to make things better, and we are not the problem.
- Love — I remastered the song I got engaged to, for my wife and the mother of my son.
- Loss and memory — this week I wrote a song about letting go of my grandfather's house and all the emotions that sparked.
What started as therapy became a letter to my son. Something I'd love to leave him about how I think today about parenthood and life. How much I love him and how I also struggle.
The Workflow: From Emotion to Album
Let me walk through the process end to end using a specific example: "Herencia", the first track on the album.
Step 1: Start with Something Real
Every night I tell my son three things: "Eres valiente, eres fuerte, eres generoso." You are brave, you are strong, you are kind. A friend mentioned this phrase to me a long time ago and it deeply resonated — these are the same values my family has always carried. That became the seed.
I also had a techno song I'd written as a joke about a year and a half ago, just to show a friend how easy it was to create melodies with these tools. That melody became the raw material.
Step 2: Generate and Iterate with Suno
I took that techno melody and remixed it into a mariachi arrangement using Suno. One thing about Suno that people don't realise: you can generate hundreds of versions of a song quite cheaply. About $20-30/month gives you thousands of generated tracks.
So I generated a mariachi version without lyrics first. Then I started layering in the words. Many versions are terrible. Others are technically fine but sound off. You keep iterating, tweaking the lyrics, adjusting the style tags, nudging the arrangement, until it clicks. You can get down to every single detail. The process is not "press a button and get a song." It's more like sculpting: you keep removing what doesn't work until what remains feels right.
Step 3: Create the Music Video
Once I was happy with the track, I thought: I might as well create a video. At Kyan we'd been using Heygen for lip-syncing from early on, constantly trying new tools. ElevenLabs gave us access to their full model suite, and I discovered that Kling 3.0 is currently state of the art for AI video — not just in quality but because it allows character consistency across scenes.
Using Gemini and Nanobanana, I could create the initial and final frame of every scene (each up to 15 seconds). Suddenly I had a visual language: my son and me, navigating life between Europe and Latin America, sometimes feeling lonely, always dealing with the challenges of being different.
I generated prompts for every scene, fed them through Kling, and assembled the video. One funny detail: the reference photos I sent for my avatar had AirPods in, and I spent an unreasonable amount of time trying to prompt the headphones out. I ended up keeping them in that video. For all the new ones, I fixed the reference images first.
Step 4: Repeat for Nine Tracks
After showing the first few songs to family and friends, several told me they had tears in their eyes. That's when I thought: maybe we're close to a full album that could actually resonate beyond my living room.
I found a rhythm. One song per week. I describe how I feel, iterate on the theme, generate variations, fine-tune the best one. For the video: 1-2 hours. For the song itself: 2-4 hours. I was building a soundtrack to my life, and I realised anyone with something worth saying could do the same.
The Full Tech Stack
Music
- Suno
- ElevenLabs
- Auxy (old melodies)
Video
- Kling AI 3.0
- Sora 2
- Veo
- Aurora
- Nanobanana
Code & Dev
- Claude Code (Opus 4.6)
- Codex
- Gemini
Marketing & Distribution
- Pomelli
- Spotify
- YouTube
- TikTok
Building the Website in an Afternoon
Once I had a solid album, the next thought was obvious: I need a website. A couple of hours with Claude Code and I had a full Next.js application running on Netlify. The site has the album, every song with lyrics, music videos, a story page, and proper Posthog analytics tracking everything. You're reading this on that site right now — leon.band.
Something unexpected happened during the video creation process. To create my digital avatar, I needed to design outfits for different scenes. Some of those generated jackets looked genuinely incredible — embroidered designs that fuse Latin American craft with streetwear. So I added a merch page with a waitlist. Technically, I now have a full product line ready to go.
Marketing with AI: The $300-600 Experiment
Finally, I used Pomelli to analyse the brand and create multiple social media campaigns targeting Latin America across TikTok, Instagram, and YouTube. The expectation is that with $300-600, the music can reach 100,000-300,000 people. I don't know the results yet, but the mere fact that a solo creator can set up multi-platform, targeted ad campaigns in hours is remarkable.
A Case Study: The Herencia Workflow, Step by Step
Start with the emotion. I wrote down the phrase I tell my son every night and what it means to me. What values I want to pass down. What I fear. What I hope.
Draft lyrics with AI assistance. I prompted around that phrase to make sure it resonates and feels raw. I used Claude to help shape the structure, but the core words and feelings were mine.
Generate the music in Suno. Took my old techno melody, set the style to mariachi, generated an instrumental version first. Then added lyrics. Generated 50+ versions. Picked the best. Fine-tuned.
Design the visual narrative. With Nanobanana and Gemini, I created start and end frames for each scene. Each scene is up to 15 seconds. I storyboarded the emotional arc visually.
Generate video with Kling 3.0. Fed the frames and prompts into Kling. The character consistency feature kept my avatar recognisable across scenes. Assembled the final cut.
Distribute. Uploaded to Spotify, YouTube, TikTok, and Instagram. Set up the Pomelli campaigns. Published on the website.
What This Means (And What It Doesn't)
Let me be clear about what I'm saying and what I'm not.
Everything but the message was AI-augmented. The emotions, the stories, the decisions about what to say and how to say it — those are entirely mine. AI didn't tell me to write a mariachi song about the values I want my son to carry. AI didn't decide that the video should show the loneliness of being an immigrant parent in Europe. AI didn't choose to leave the AirPods in the first video because honestly, it kind of works.
What AI did was remove the barriers between having something to say and being able to say it in the medium I wanted. I couldn't afford studio time. I can't play mariachi instruments. I don't know After Effects. I'm not a web designer. None of that mattered.
We're reaching a point where you can make your own movies, your own music, your own soundtrack. What I'm doing is not trivial — you still need to make hundreds of decisions and understand what each tool can and can't do. But it's a massive step forward, and we're not far from the tools handling everything together seamlessly.
The Bigger Picture: Augmented Teams
At Kyan Health, we operate under these principles every day. What I did with music, our team does with product, content, research, and operations. We're constantly evaluating new tools, testing capabilities, and figuring out how to integrate them into real workflows. The result is that a small team can output work that would have required a much larger one just two years ago.
I'm not sharing the inner workings of how we work at Kyan. But I'm sharing this side project as proof of what we've learned: anyone on our team could do a project like this. Not because they're all musicians, but because they've all learned how to be augmented.
This album is a prequel of what's coming for a lot of people. If you have great taste and something worth saying, you will soon be able to create things that were previously reserved for teams with six-figure budgets. The tools are here. The cost is negligible. The only bottleneck is having a message that matters.
A Note on the Music Itself
The album is called Paternidad. It's nine songs about being a father, an immigrant, a founder, a husband, and a person trying to figure it all out in real time. Mariachi meets emo-punk. Spanish lyrics. A lot of raw emotion.
Some tracks that might resonate:
- Herencia — The values I want to pass to my son. Be brave, be strong, be kind.
- El Peso — The weight of trying to be everything, everywhere, for everyone.
- Cuando Hables de Mi en Terapia — When you talk about me in therapy. Because every parent knows they'll end up there.
- Corazon de Leon — The title track. A lion's heart. Fierce love.
- Digame que Si — A remastered version of the song I got engaged to.
- Campo Minado — Navigating parenthood and life like walking through a minefield.
- Donde Dolia — Where it hurt. Letting go of my grandfather's house.
- Porque Si y Porque No — The contradictions of being a parent. Because yes and because no.
- Escala de Gris — Life in grayscale. Nothing is black and white once you have a child.
I realised that I can make one song per week just describing how I feel. Then I iterate on that theme, create a video in a couple of hours, and suddenly I have a soundtrack to my life. And so can you.
Because if you have something worth saying, you can now get your message across in ways that were impossible before. The art is in the message. Everything else is becoming infrastructure.
Listen to Paternidad
Nine songs. Nine truths. Made with a lot of love and a little bit of AI.
Listen on the Site Watch the VideosIgnacio Leonhardt is the CTO of Kyan Health and makes music as Leon. Follow him on leon.band or any platform where you listen to music.