March 2026 · By Jesse Meria
I run a cafe in Charlevoix, Michigan. Every song that plays in the space was generated by AI. The prompts come from HookGenius. The generation happens in Suno. The best tracks end up in Puana, where other businesses can use them too. This is how the whole pipeline works.
Background music for a business is a solved problem, technically. You pay for a licensing service. They give you playlists. You press play. I did this for years. It worked.
But it never felt right. A cafe has a personality. Mine has a specific one. The music should reflect that, and curated playlists from a licensing service reflect the taste of whoever built the playlist, not the taste of the space.
I wanted custom music. Music that matched the specific vibe of a morning pour-over crowd. Music that shifted for the afternoon. Music that set a mood for summer evenings. Not someone else's idea of "cafe music" — my idea of it.
Suno made custom music possible. AI music generation went from a toy to a legitimate production tool in about a year. By late 2024, the output quality was high enough that I could generate tracks and play them in the cafe without anyone noticing they were AI-generated.
But generating good tracks consistently required good prompts. And writing good prompts for Suno is a skill. One that took me months to develop through trial and error. Specific genre tags. Precise vocal direction. Structure tags in the right places. Production descriptors that actually map to Suno's model. Character limits that matter more than most people realize.
I codified everything I learned into a tool. That tool became HookGenius.
Step 1: Generate Prompts
I use HookGenius to generate style prompts and lyrics for whatever I need. Morning session? Smooth jazz with warm saxophone, brushed drums, instrumental. Afternoon? Indie folk, acoustic guitar, gentle vocals. I describe the vibe, HookGenius translates it into Suno-optimized prompts with the right tags, structure, and vocal direction.
Step 2: Generate Tracks
I paste the HookGenius output into Suno and generate. Usually 3-5 variations per prompt. The HookGenius prompts are specific enough that most variations are in the right neighborhood. I pick the best one. Sometimes I generate more if nothing hits.
Step 3: Curate
Not every track makes the cut. I keep roughly 1 in 5. The filter is simple: would I play this in the cafe without thinking about it? If I have to convince myself it's good enough, it's not. The bar is "sounds professional, fits the space, no AI artifacts."
Step 4: Play
The tracks that pass go into rotation. I have playlists by time of day and season. Morning jazz. Afternoon acoustic. Evening ambient. Summer gets brighter genres. Winter gets warmer ones. The cafe soundtrack is entirely custom and entirely AI-generated.
Step 5: Share
The best tracks from my curation process end up in Puana. It's a curated AI music library for businesses. Other cafes, restaurants, retail shops, co-working spaces — anyone who needs background music can browse by mood, genre, and energy level. Every track in Puana was generated with HookGenius prompts and passed my personal quality filter.
Less than what I was paying for a commercial music license. Significantly less.
A Suno subscription covers the generation. HookGenius credits cover the prompting. The curation is my time, but it's time I enjoy spending — listening to music and picking what fits my space. Compared to the old model of paying a monthly licensing fee for playlists someone else built, this is cheaper and produces better results.
For businesses that don't want to generate their own music, Puana offers the output directly. Pre-curated, pre-filtered, ready to play. The generation work is already done.
Nobody has ever said "is this AI music?" Customers ask what the playlist is. They ask if I can share it. That's the bar: people don't notice it's AI. They just notice the music is good.
This is important because it proves the output quality argument. AI music doesn't sound like AI music anymore. Not if the prompts are right. The tracks in my cafe are indistinguishable from human-produced independent music. Some of them are better than what the licensing service was providing.
Background music is a commodity expense for most businesses. A monthly fee for a service you don't think about. But music shapes the experience of a space more than most owners realize. The right music makes people stay longer, buy more, and come back.
Custom music used to mean hiring a composer or licensing specific tracks. Both are expensive. AI generation changes the economics. You can create a soundtrack that's tailored to your exact space, your exact brand, your exact customers. And you can do it for less than a generic licensing service costs.
If you want to try this pipeline yourself:
This isn't just about background music. It's about what happens when AI tools mature to the point where creative output is genuinely professional. Music was one of the first domains to get there. The tools exist now to create, curate, and distribute AI music at a quality level that serves real businesses.
HookGenius is the creation layer. Puana is the distribution layer. Together, they represent a pipeline that turns an idea ("I want jazz for my cafe") into a curated library of professional tracks. No musicians hired. No licensing negotiated. No compromises on quality.
That's the future of business music. And it's already running in a cafe in northern Michigan.
5 free generations on HookGenius. See what AI prompting can do for your music.
Try HookGenius Free →Puana has thousands of AI-curated tracks ready for your business. Browse by genre, mood, and energy.
Explore Puana →