Built This Week/Builds

Builds

AI DJ for Retail Stores

Our AI retail DJ swaps between five custom AI-generated songs, 90 to 130 BPM, as a store fills up and empties out, so the music follows the crowd instead of a generic playlist.

For Episode 21 (November 14, 2025), Jordan built a "retail mood AI DJ mixer" from an idea of Sam's. It's a store DJ that plays five custom AI-generated songs, from 90 BPM up to 130 and back down to 100, matched to how many shoppers are in the store. We built it with our guest's company, Standard AI, in mind. The episode didn't say how long it took.

Why we built it

Jordan's starting point at 04:14: most clothing stores play whatever is popular, and it doesn't change whether the store is busy or empty, morning or night, raining or sunny. Nobody is using data to pick the music.

So we thought we could use AI and computer vision to kinda build a DJ that would play different music throughout the day depending on the occupancy of the store.

— Jordan Metzner, Episode 21

Sam's thesis was that beats per minute might correlate with sales, or the opposite: a candle shop may not want 130 BPM, but a sunglasses store might. Either way, Standard AI's computer vision could measure it. The inspiration on the product side was Spotify's AI DJ, which Jordan finds "sometimes cool and sometimes annoying."

The stack

  • An AI music generator for the five tracks. Jordan said he loves Suno and has tried Udio and ElevenLabs' APIs, but didn't say which one he used here.
  • A demo app that simulates occupancy and switches tracks. The episode didn't cover how it was coded.
  • Standard AI's computer vision (not connected in the demo): the source of real occupancy and shopper behavior data.

How we built it, step by step

  1. Generate five songs at different tempos. They run from 90 BPM up to 130 and back to 100. The opening track includes a spoken welcome for shoppers.
  2. Map each track to an occupancy level. Opening mode, then busier and busier up to a full store, then back down to closing mode.
  3. Build a demo that walks through the day. On air, the simulation starts at an 11 AM beat, the "party gets higher and higher" as occupancy climbs, then the music comes back down to chill for closing.
So they go from 90 BPM all the way up to one thirty and back to a 100.

— Jordan Metzner, Episode 21

How it turned out

It's a working concept demo, not a deployed product. Standard AI CEO Angie Westbrook liked it and at 07:53 said she'd want to put it in one of their stores to see whether the upbeat-music hypothesis holds. Her reasoning: you can't always see an instant reaction in sales data, but Standard AI tracks what shoppers actually do and whether they browse more.

What we'd do differently

The next version would test tracks against each other. Jordan's example at 08:16: if disco works best in the afternoon, compare individual disco songs and keep iterating until one maximizes sales, then repeat by store type. Angie added personalization: changing the music for shoppers in a certain aisle, or learning preferences over time.

Jordan also pitched store-specific holiday music, so malls stop playing the same Mariah Carey song on repeat. For more on how stores use computer vision, see AI in retail and e-commerce, and for our other music build, the Strudel AI DJ mixer. More music tools are on AI in music.

FAQ

What is an AI DJ for retail stores?

Our version plays custom AI-generated music that changes with how many shoppers are in the store: five tracks mapped from opening mode, through a packed store, to closing mode.

Does store music BPM affect sales?

That's the hypothesis we wanted to test, and we haven't proven it. Sam thought faster music might drive sales in some stores but not others, like a candle shop. Angie Westbrook of Standard AI said sales data alone can't show an immediate reaction, but tracking shopper behavior could.

How do you use computer vision with store music?

Computer vision counts how many shoppers are in the store and the DJ picks the matching track. Standard AI turns security-camera footage into shopper behavior data, which is what would let a store measure whether the music changes browsing.

What AI music tool did you use?

Jordan said he loves generating music with Suno and has tried Udio and ElevenLabs' APIs, but the episode didn't say which one made these five tracks.