🎵 The Playlist Is Not the Set

AI can hand you 200 songs in 10 seconds. It can’t tell you which three will get guests off their butts. Only real performance experience can.

The Playlist Is Not the Set

A DJ can now type “high-energy wedding reception, mixed ages, couple in their late-twenties, couple likes 2000s pop” into a chat window and get a complete playlist back before a complete song plays.

The playlist will be okay. It might even be good.

But that is exactly the problem.

The new way to be average

Amanda Natividad, Chief Evangelist at SparkToro and co-author of Zero Click Marketing, recently wrote about, of all things, a Pokémon building game.

Her argument should resonate with every wedding DJ. Tools that remove technical barriers are great. Tools that make the creative decisions for you are something else entirely. The joy, and the skill, live in imagining something. Hitting obstacles, changing your mind, and picking somewhere better than you planned is all part of the game.

She also pointed to a 2024 study in Science Advances where writers using AI produced stories rated as more creative. However, those stories ended up more similar to each other. Better individually. Flatter collectively.

Now picture that across the wedding industry. Hundreds of DJs prompting the same tools with the same couple questionnaires. Every playlist is competent. Every playlist looks like every other playlist.

The couple can’t tell them apart, because there’s nothing to tell apart.

That’s not a technology problem. It’s a taste problem. Taste doesn’t come from a prompt!

I use AI. Here’s what it can’t do.

Full disclosure: I use AI every week. Not to write this newsletter, but to edit it. I ask it what to cut, what’s missing, and what would tighten the argument. It’s a good second set of eyes.

I’ve also asked it to build wedding playlists. The results were fair at best.

Safe picks, obvious picks. The lists were the songs anyone would name in the first 30 seconds of thinking about it.

Nothing that reflected what actually happens when 150 people with three generations of musical memory hit a dance floor at the same time.

That gap isn’t a bug the next model will fix. It’s the difference between knowing about songs and knowing what songs do in a room.

What a playlist actually is

Here’s the belief I want to replace:

A lot of DJs think the playlist is the work. Build the right list, press play in the right order, done.

The playlist is a guess made in advance. The set is what happens when the DJ reads the room.

Every DJ who has been doing this long enough knows the moment. The list says it’s time for the throwback run. The floor says otherwise.

Aunt Linda’s table is up, the groomsmen just sat down. The bride’s college friends are hovering near the bar waiting for permission. No playlist knows that. You know that, because you’re looking at it.

I remember a moment when I wanted to switch things up and go in a different direction. Knowing the couple and the crowd, I needed to cross between rock and Latin. I chose “Smooth” by Santana and Rob Thomas. It was a perfect transition between the genres, and guests loved it.

The three things AI can’t do for you

If you analyze AI against my 5 layers of Music Architecture, the gap becomes very clear.

1. Discovery that’s been tested (the Certainty Crate). Your Certainty Crate wasn’t handed to you. You built it by finding songs, playing them for real crowds, and watching what happened. Every trusted track earned its spot on a dance floor, not on a list. AI can suggest songs. Only real performance experience tells you which suggestions to trust.

2. The blend (Transitions & Cohesion). A list of great songs is not a set. Whether “Mr. Brightside” lands as a peak or a closing track depends on what came before it and how you got there. Keys, tempos, the energy of the outro, whether the drop gives the floor a breath or a shove. You learn that by prepping mixes, rehearsing transitions, and having a few crash-and-burns, as they say. There are no shortcuts.

3. The room (Moment Matching and Energy Architecture). Reading a room is real-time recognition built from hundreds of events. Who’s dancing, who’s drifting, who’s about to leave, which generation needs a song right now. The playlist is a template. The room is the pattern. Only one of them is happening live.

Where AI belongs

I’m not telling anyone to delete the tools. I use them, and you should too.

Use them the way I do: as an assistant, not the author. Tagging BPM and key. Cross-checking a do-not-play list against a couple’s must-plays. Building a first-draft Crowd Coverage map so you can see which generations are thin. Pulling release dates.

Let AI handle the logistics. Keep the decisions.

The moment you outsource the decisions, you stop doing the thing that builds your judgment. Your judgment is the product. It’s the reason a couple hires you instead of renting speakers.

The premium is shifting

As polished-but-generic playlists get cheaper to produce, the DJ who can prove discernment becomes more valuable, not less.

The couple can’t articulate it, but they feel it. The floor that never empties. The song that made Grandpa stand up. The transition that got a cheer.

None of that came from a prompt. It came from real performance experience, one room at a time.

The takeaway: AI can build the playlist. Only you can build the set.

Thanks for reading.

Cheers,

Matthew Campbell

Wedding MusicLetter

P.S. Check out The Wedding Music Canon - the best wedding songs from the 1950s-2010s.