Hi,

A while ago, I wrote a newsletter article called The Zombie Music Industry - it analyzed how the evolution of AI on the internet could affect the music industry in the form of AI-generated traffic.

In recent weeks, we’ve seen the explosion of AI Agents launching that are designed for the everyday person. As these platforms grow, they have the potential to change user behavior and affect the future landscape of music discovery, as well as how we live our day-to-day lives.

Let's get into it below.

———————————————————————————

This newsletter highlights:

  • The future of music discovery

Let's dive in ⬇️

———————————————————————————

Before diving into this specific newsletter edition, it’s worth noting that it does admittedly focus mostly on AI. That said, I am personally not a fan of generative AI (using AI to create music, artwork, etc) and solely use AI to help take care of my administrative tasks, work smarter, and faster. Staying ahead of the latest trends has allowed me to not only build my own tools but increase the value I can provide to the artists and creators I work with. You won’t catch me advocating for AI-generated music.

That said, for those that may not be familiar, the last few weeks in the AI world have been pretty explosive. Many of the big companies have been launching their own personal agent products. For example:

  • SpaceX launched GrokBot

  • Meta launched Muse

  • Anthropic launched Claude Duet

  • OpenAI launched Dot

These agent platforms essentially act as teammates and assistants that can help execute tasks. You can connect other platforms to them (for example: email, calendar, Google Drive, etc) and the Agent can interface and use those products without having to log in.

For example, I use one of the tools mentioned above and created +5 different AI agents with it. I have my email, Google Calendar, Google Drive, and more connected to it. Without overexaggerating…it has legitimately changed the way that I work.

For context, here are some of the things it does for me:

  • Scans my emails every morning and tells me what’s most important and urgent to respond to

  • The night before meetings, it gives me a quick brief on those meetings for the upcoming day

  • Scans the social media accounts of artists I’m working with daily, rips new content, and puts it into a Google Drive folder for me

  • Scans my email for mention of paying vendors / budget and automatically adds payments to a running Google Sheets so I don’t have to manually update it

It’s almost like having a personal assistant that almost never misses anything, is proactive, and has an insane level of attention to detail. What’s essentially happening is that all of these AI companies are trying to create the best product for the least-technical person and own the category.

Whichever one of the companies mentioned above develops the easiest product to use for the person who has no idea what an AI agent is and provide them the most value, will likely get the most amount of subscribers and users. The goal, as it seems, is to make these Agent platforms a one-stop shop for everything you would need in your day-to-day. For example, in theory, I could do all of the below from my personal agent:

  • Read + send email

  • Scan social media

  • Look at my calendar

  • Edit Google Docs or update docs

  • Order food

  • RSVP for events

  • ec

There’s a method to their madness…they are trying to make these platforms “portals” to the rest of the internet.

A friend shared with me that on a recent episode of the All-In podcast, one of the hosts named Chamath made a similiar point: AI agents (ChatGPT, Claude, Grok, etc) are trying to become the one place you go to for everything. As they do, the apps they connect to become "headless": meaning that they still run in the background, but fewer people actually log in anymore because you can just ask the agent.

It got me thinking…as this develops, how could this impact the music industry and music discovery? In a way, we’re already seeing the early signs. For example, when you ask ChatGPT for music, it can pull up Spotify right inside the chat and use your conversation to determine what it recommends.

As of March of this year, we can actually hear audio previews on Spotify through ChatGPT without logging in. It makes me wonder…does this potentially create a future where less people actually log into Spotify and more people just use their Agent to recommend music? If so, what does that mean for music discovery? How likely is that?

How I think it plays out

Here’s how I think it plays out…

My guess is that Agents absorb most of the “utility'“ work: for example, dashboards, admin panels, checking tour dates, reordering merch, etc. Those are essentially tasks, and tasks can get delegated.

Being entertained is different though - I don’t think people will ask an agent to scroll TikTok for them, because scrolling and consuming content is the point of the platform. Entertainment will likely stay in the apps themselves while utility moves to agents.

It’ll be interesting to see how this plays out because Spotify makes its money when you're inside Spotify, which is why playback in that ChatGPT integration still sends you back into the Spotify app.

YouTube is taking a slightly similiar approach. Last month, it introduced its Ask YouTube feature to everyone in the US across web, mobile and TV. You don't need to search "Larry June interview" anymore - in theory you could ask "what's a good breakdown of how independent rappers make money?" and get a written answer with videos, Shorts and the exact timestamped clips, then follow-up questions to dig deeper.

If this trend continues, platforms basically evolve from giving specific answers to coming up with their own answers based on a set of criteria or questions.

We've seen this before

Remember when everyone was buying Amazon Echos? Music actually became one of the main things people looked for with smart speakers.

Labels also found that smart speakers often couldn't find artists with dollar signs or unusual spellings in their names. (Curren$y must have been living in Alexa hell for years lol.)

The Recommendation Economy

There’s a difference this time around though…while Alexa could only handle "play Drake”, Agents can handle an ask like "something like Larry June but jazzier, for a Sunday morning drive."

It’s interesting to look at what changed in that request - it doesn't name anyone, it just gives a criteria. Ask YouTube works the same way: you search by criteria, not by people. It’s a real discovery opportunity, but only for artists the agent knows how to describe.

A term for this could be the Recommendation Economy - when discovery shifts from people browsing to these platforms recommending music for them, the goal could in theory transition from getting on a playlist to becoming the recommended answer. Agents build their answers from what the internet says about you: press, interviews, Reddit threads, Genius pages, video titles and descriptions, how fans describe your sound.

Now…there is a valid argument to be made that nobody is going to ask AI for music recommendations. Most people aren't yet, which is true. But YouTube just launched the ask feature in the US, so the habit may form whether we're ready or not. My guess is that by 2030, maybe 10–20% of "what should I listen to" moments start with an ask. That's not most of discovery, but it could be the fastest-growing piece, and the artists who show up there early get the compounding and discovery.

What to do about it

1. Audit what the agents say about you. Open ChatGPT, Claude, or Gemini etc and paste this:

❝

"Who is [artist name] and what do they sound like? Recommend 5 artists similar to them, and tell me which of their songs to start with."

Then flip it:

❝

"Recommend 5 up-and-coming [genre] artists from [city] who sound like [bigger comparable artist]."

If you're missing from the second answer, or the first one is wrong, that could be your to-do list. It shows you the story the internet is telling about you and the gaps you need to fill.

2. Pick one sentence for your sound and use it everywhere. DSP bios, press one-sheets, interview answers, your website. Consistent language gets repeated, and repeated language is what agents learn from.

3. Describe your videos the way people ask questions. Most artist uploads are titled "Artist – Song (Official Video)" and nothing else. Add the context someone would actually search for: the vibe, the setting, the story behind it. "Late-night Houston R&B, recorded in one take in my mom's garage" gives an AI something to work with. A title alone doesn't.

4. Clean up your name and metadata. Same spelling everywhere, credits filled in, and a Genius page that actually exists.

5. Own the relationship. This is the big one. If the platforms turn into a recommendation layer, the email list, the text list and the community are the only places where nothing sits between you and the fan. (Agents will summarize emails too, so write a newsletter people want to read all the way through, not one they skim.)

6. Build IRL. The more of the internet that gets automated, the more important IRL will become.

Overall

For context, I don't think apps become irrelevant and I don’t think people completely stop using Spotify, YouTube, Apple Music, etc to discover music. I just think that as the only mechanism for discovery slowly evolves.

Thanks for reading, until next time.

P.S. If you're an independent artist, manager, or music company trying to get a release off the ground in the next few months, I'd love to help. Tell me what you're working on HERE.

———————————————————————————

What'd you think of this week's newsletter?

Your feedback goes a long way.

Login or Subscribe to participate

Reply

Avatar

or to participate