I Made a UK Rap Song With AI After Losing My Father (2026) Meta Description: How I turned grief into a UK rap track and cinematic music video using AI tools, with zero music or film background. The full process, free guide included.
I Made a UK Rap Song With AI After Losing My Father — Here's Exactly How
Published: August 18, 2026
Quick Answer: Yes, you can produce a full UK rap track with an AI-created, London-style cinematic music video without any music production, singing, or filmmaking background. I did it as a tribute to my late father and my mother — writing the concept, generating the vocals and instrumental with AI music tools, building the visual scenes with AI video generation, and directing the whole creative arc myself. The song is called "My Diamond." Below is the real process, the tools, the mistakes, and a free step-by-step PDF guide.
The Night I Decided Grief Wasn't Going to Just Sit There
I lost my father. That sentence still feels strange to type, even months later. There's a specific kind of quiet that shows up after — not sadness exactly, more like the room getting bigger and you getting smaller in it. My mother was the one who stayed. She didn't fix the grief, she just stood next to it with me, every single day, while I tried to figure out what to do with a business, a laptop, and a future that suddenly felt heavier than before.
I'm not a musician. I've never written a verse in my life. I don't play an instrument, I can't sing, and I definitely can't direct a music video with cars and motorbikes weaving through London streets. But I had this pressure in my chest that needed somewhere to go, and scrolling past another "AI can now do X" headline, something clicked: what if I pointed all of that at grief instead of at another SEO article?
That's the honest origin of "My Diamond." Not a content strategy. Not a growth hack. A guy at 2am in Casablanca, missing his father, grateful for his mother, with a laptop full of AI tools and no idea if any of this would actually work.
Why "Just Use AI Music Tools" Is Where Most People Get Stuck
Here's the part nobody tells you when they say "AI can make a whole song now." Technically, sure. Practically, it falls apart fast if you don't understand the layers involved, and most beginners try to do it in the wrong order.
The instinct is to open one AI music generator, type "sad UK rap about losing someone," hit generate, and expect a finished track. What you get instead is generic. Flat drums, vocals with zero emotional pocket, lyrics that sound like they were written by someone who read about grief in a textbook instead of living it. The tool isn't the problem. The prompt is.
The real complication is this: AI music and video generators are extremely good at execution and extremely bad at meaning. They can build you a beat, a flow, a visual scene — but they cannot tell you what the song is about at a level that matters. That part has to come from you, fully formed, before you touch a single tool. If you skip that step, you end up with content. If you don't skip it, you end up with something people actually feel.
So before I opened any AI tool, I wrote down, in plain words, what "My Diamond" needed to say:
- Pain turned into fuel, not into paralysis
- A mother who becomes the anchor when everything else moves
- London as the visual backdrop — cinematic, moving, restless, like the feeling itself
- No self-pity in the lyrics. Ambition instead. Motion instead of stillness
Only after that did the tool selection even matter.
What I Actually Used, Step by Step
I'm not going to pretend this was one clean pipeline, because it wasn't. It was trial, error, re-generation, and a lot of late nights. Here's the real sequence:
1. The concept and lyrics. I wrote the emotional skeleton myself — the story of loss, my mother, the "diamond" metaphor (pressure creates diamonds, and that's what grief did to me). I used an AI language model as a collaborator to help shape raw feeling into UK rap cadence and slang, going back and forth line by line until it sounded like something a real London MC would actually say, not something a machine assembled from rhyme dictionaries.
2. The instrumental and vocals. I used an AI music generation tool to build the beat and vocal performance, feeding it the lyrics along with detailed style direction — UK drill-adjacent rap, moody low end, minor key, London energy. This step took the most regeneration. Early versions sounded like karaoke. It took specific, almost obsessive prompt tweaking around vocal tone, delivery pace, and emotional weight before it stopped sounding like a demo and started sounding like a track.
3. The visual concept and scenes. For the music video, I used AI video generation to build the cinematic sequences — London streets, a car scene, motorbike shots, atmosphere that matched the emotional arc of the song rather than just looking cool for its own sake. Every scene was chosen to mean something: motion representing forward momentum, the city representing scale and ambition, the lighting representing the tension between dark and light that grief actually feels like.
4. Creative direction and editing. This is the part people skip when they talk about "AI making content," and it's the part that actually took the most effort. I storyboarded the shot order, decided pacing against the song's structure, picked which AI-generated clips earned a place in the final cut and which didn't. AI generated the raw material. I directed it into a story.
5. Distribution. Once the track and video were done, I published across LinkedIn, X, and YouTube Shorts, each with its own framing, because a platform-native post outperforms a copy-pasted one every time. LinkedIn got the personal story format you read above. YouTube Shorts got the visual hook first, story second.
I also documented the entire workflow into a step-by-step PDF, tool by tool, prompt by prompt, because I built Fikrago on the principle that if I figure something out from zero, I write it down so the next person doesn't have to start from zero too.
The Mistakes That Almost Killed the Project
I want to be honest about the parts that didn't work, because every "how I built this" post skips them and that's exactly why people give up on their own version too early.
The first full draft of the instrumental was wrong in a way I couldn't articulate at first. Technically it was fine — decent tempo, decent mix — but it felt like elevator music wearing a UK rap costume. It took me a full evening to realize the problem wasn't the beat, it was that I'd described the genre to the AI tool instead of describing the feeling. "UK drill-influenced rap" gave me generic UK drill-influenced rap. It wasn't until I started prompting with emotional and physical detail — tight chest, restless energy, a beat that sounds like pacing back and forth at 3am — that the output started matching what was actually in my head.
The vocals were worse at first. Early takes had the right words but the wrong weight. Grief doesn't sound aggressive, and my first few generations came out sounding almost angry, which wasn't the story. I had to explicitly redirect the tool toward restraint — a delivery that holds back instead of pushing forward, because that's closer to what real grief sounds like when someone's trying to stay composed in public. That one adjustment changed the entire character of the track.
The video had its own failure mode. My first batch of AI-generated London scenes looked like stock footage — technically London, emotionally nowhere. Wide shots of streets with no person in them, no sense of motion tied to a story. I had to start treating every generated clip like a sentence in a paragraph, not a standalone image. Does this shot come before or after the one that shows momentum building? Does the lighting shift track with the emotional shift in the lyrics at that timestamp? Once I started directing scene-to-scene instead of clip-by-clip, the video stopped looking like a demo reel and started looking like a story.
None of these were tool failures. They were direction failures on my end, and fixing them was less about better software and more about getting more specific about what I actually meant.
What I'd Tell Someone Starting From Zero, Like I Did
If you're reading this because you're sitting with something heavy and wondering if turning it into something creative is even worth trying — a few things I'd say, tool-agnostic:
Don't start with the software. Start with one sentence that's completely true about what you're carrying. Everything downstream — the lyrics, the beat direction, the visual choices — should trace back to that one sentence. If a generated output doesn't serve that sentence, cut it, no matter how good it looks on its own.
Expect to regenerate far more than you'd like. I lost count of how many vocal takes and video clips never made the final cut. That's not wasted effort, that's the actual process. The tool isn't failing you when the fifth version still isn't right — you're still refining what you're trying to say, and the AI is just reflecting that back at you.
Separate the emotional work from the technical work, even though they happen at the same desk. There were nights I had to stop, step away, and just sit with missing my father before I could come back and make a good creative decision about the video. Trying to force both at once produces work that feels hollow, and people can tell.
And don't skip documenting it as you go. I wrote down every prompt, every tool, every dead end while it was happening, which is the only reason the step-by-step PDF guide exists now. If you wait until after you're "done" to document the process, you'll forget the specific decisions that actually mattered.
The Part That Isn't About Tools At All
Here's what I didn't expect. Somewhere in the middle of adjusting prompts and re-rendering vocal takes at 1am, the process itself became part of the grieving. I wasn't avoiding the feeling by staying busy — I was translating it into something that would outlast the moment it came from. Every regenerated take was, in a strange way, another attempt to get the feeling right, not just the audio.
My mother watched the final video before I posted it. She didn't say much. She didn't need to. That's the diamond — not the metaphor in the lyrics, the actual woman who stayed present through the hardest stretch of my life while I figured out, piece by piece, how to keep building instead of stopping.
If you've lost someone, or you're carrying something heavy while everyone around you assumes you're fine because you're still showing up to work — I'm not going to tell you AI tools will fix that. They won't. But they gave me one honest way to put the weight somewhere outside my own chest, and that mattered more than I expected it to.
What This Means If You Want to Try Something Similar
You don't need a music background, a film background, or a big budget. You need three things, in this order: a real story that's actually yours, the discipline to direct AI tools instead of letting them decide for you, and the patience to regenerate a scene or a vocal take fifteen times until it stops sounding like a demo.
Most people fail at step two. They treat AI generation as a vending machine — put in a prompt, get out a finished product — instead of as a collaborator that needs direction, correction, and a clear creative vision behind every prompt. The people getting genuinely good results right now are the ones directing, not the ones delegating.
FAQ
Do I need any music production experience to make a song like this with AI? No. I had zero background in music production, singing, or songwriting before this project. The skill that matters more is being able to describe what you want clearly and being willing to regenerate outputs until they match the feeling you're going for.
What AI tools are involved in a project like this? Broadly, you need an AI language model for lyric development, an AI music generation tool for the instrumental and vocals, and an AI video generation tool for the visual scenes. The specific tools matter less than the workflow and creative direction behind them, which is what the full PDF guide walks through.
How long did the whole project take from idea to published video? Realistically, spread across nights and weekends around running Fikrago, it took a few weeks — not because any single step is slow, but because getting the emotional accuracy right in the vocals and visuals required a lot of regeneration and revision.
Is this something brands or businesses could use, not just personal storytelling? Yes. The same workflow — concept first, AI execution second, human direction throughout — applies to brand storytelling, product launch videos, or founder-story content. The emotional core just needs to be replaced with whatever story is actually true for that brand or founder.
Where can I watch the video and get the full process guide? You can watch "My Diamond" on LinkedIn, and I've put together a complete step-by-step PDF covering every tool, prompt, and decision in the process — link below.
Related on Fikrago:
Written by Ayoub Zinani, Founder of Fikrago. Google Certified in AI Essentials, Digital Marketing & E-commerce, and Cybersecurity. Ayoub builds AI-assisted tools, content, and creative projects from zero, documenting the entire process for anyone starting with nothing but a laptop and a reason to keep going.
Watch the video: My Diamond — LinkedIn
Want the full process? Grab the step-by-step "My Diamond" creation guide or explore more AI-built tools and templates at fikrago.com/p/tools.html.