After the Track: How Nik McFly Turns AI Music Into a Real Release Business
Gary WhittakerAfter the Track: How Nik McFly Turns AI Music Into a Real Release Business
The prompt may be the easy part. The harder work starts when a song has to become a release, an artist identity, a set of records, a marketing decision and eventually a business.
Nik McFly first reached out to me because he was researching the release problems AI music creators hit before upload: rights, metadata, distributor rules, Content ID and platform trust.
That conversation quickly became more interesting than a checklist.
Nik is a songwriter and music entrepreneur who creates with tools including Suno, works with his wife on music, co-founded and runs an AI-native music label, runs AI Music Events, and teaches creators what happens after an AI-assisted track exists.
In his written responses for this Creator Spotlight, Nik reported that the shared catalog he and his wife built has grown to 80+ AI-music releases, more than 20 million streams, and $65,000 in revenue received as of September 2026. He was also careful to distinguish money actually received from reported royalties and visible streaming numbers.
The useful story is not that AI made music easier to generate. It is what had to be built after generation stopped.
Note: catalog figures in this article are based on Nik McFly’s written interview responses and are presented as his reported results.
“The prompt was the easy part.”
That is one of the most useful ideas in Nik’s answers.
AI can dramatically increase the number of songs a creator can generate. But that creates another problem: more music means more decisions.
A track can sound finished and still have the wrong chorus. It can be technically usable and still feel wrong for the artist. It can duplicate something already released. And even when the song is worth keeping, it still has to connect with artwork, distribution, credits, rights records, promotion, payments and a release schedule.
Nik described the operational pressure that appeared when several songs came out close together. The team could no longer easily tell which releases were actually bringing listeners back. Marketing focus became diluted. Each song had less room for its own story.
More output is not automatically more progress. At some point, the constraint moves from generation to selection, positioning, release operations and attention.
How does he decide a track is actually worth releasing?
Nik’s answer was refreshingly non-technical.
After listening to many generated versions, he looks for a physical response: his body starts moving before he has to think about it.
That is the point where he stops generating and starts developing.
The team then works on the sound, vocals and mastering while trying to preserve the feeling that made the track work in the first place.
This matters because AI creators can get trapped in endless regeneration. A technically “better” version is not always a stronger song. At some point, the creator has to recognize the useful signal and commit.
Artist positioning comes before scale
Nik uses a simple sentence to help define an artist:
“People listen to my music when they want to…”
The answer might be to feel confident, calm down, dance, process a difficult emotion or put something into words.
That sentence becomes a filter.
It helps determine what songs belong, how the lyrics should feel, what the visual world should communicate and, just as importantly, what should be left out.
For AI-assisted music, this is one place where human direction becomes obvious. A model can generate possibilities. The creator still has to decide what the artist is for.
Twenty million streams did not remove the need for better records
One of Nik’s strongest operational lessons is that creators should not treat a visible stream count, a royalty report and a payment as the same event.
He described a case where a track showed millions of plays in platform analytics while the distributor had not yet reported royalties for the same period.
That forced the team to trace which distributor handled the release and which reporting period those streams belonged to.
When the audience actually streamed the track.
When the distributor or platform reports income.
When money is actually received.
That distinction sounds basic until a catalog becomes large enough that multiple releases, distributors, reporting periods and payments overlap.
The operating side of the catalog
Nik describes his own role as leading strategy, partnerships and the systems behind the catalog.
That includes connecting release records with financial reports, tracking costs, understanding what individual tracks earn, reducing repetitive work and using the resulting information to decide where time and budget should go next.
It is a useful reminder that an AI-native music operation is still an operation.
Files need names. Contributions need records. Permissions need documentation. Versions need to be tracked. Artwork has to match the correct release. Metadata has to stay consistent. Costs need to be compared with results.
None of that disappears because the first demo was created quickly.
Why he built AI Music Events
Nik kept encountering the same problem: information creators needed was scattered across many websites.
Which AI-music events were actually happening? Where could an artist apply? Which rules affected a particular release? What did a distributor currently allow?
That led to AI Music Events.
It also led to Can I Release This?, a free tool designed to check a release plan against the public rules it covers, compare distributors, show sources and flag areas that still need checking.
Nik’s standard for information is straightforward: separate a claim from a verified fact.
His three questions for readers are equally useful:
Where did this information come from?
Is it still current?
Does it apply to what I want to do?
If release operations have been an afterthought, start with one song
Nik does not recommend fixing everything at once.
His suggested starting point is to finish one track and choose a distributor that accepts the type of music you make. Check the distributor’s current requirements before uploading.
Then take that one song through the full process:
Prepare the audio.
Prepare the artwork.
Document the credits.
Choose the release date.
Check the current distributor requirements.
Confirm the song lands on the correct artist pages.
Keep the files and release records together as you go.
Then use what you learned to prepare the next release.
That is a much more durable path than trying to build a giant release machine before you have completed the process once.
From personal catalog to AI-native label
Nik says he and his wife have now launched an AI-native music label and brought their first artist onboard.
His definition of AI-native is important: AI tools are part of how the work is created and operated from the beginning, while people remain responsible for the creative and business decisions.
The next challenge is applying what the team learned from its own catalog to someone else’s music.
That means helping artists develop a clear sound, prepare releases and reach listeners — and finding out which parts of the original system actually transfer.
That is a more demanding test than simply producing another track internally.
What Nik is building next
For creators who already have promising AI demos and want to turn them into finished releases, Nik pointed me toward his upcoming Maven course:
Suno & GPT-6 Astra: The AI Music Production Lab
The course is designed around song development, production in FL Studio, mixing, release preparation, visuals and content. Nik teaches creative planning, AI-assistant workflows and release operations, with Vladimir Krasovitskiy teaching production, mixing and mastering, and Bally Sol teaching visual storytelling and content.
View the Maven course →He also shared After the Track, a practical release-planning resource with editable templates for schedules, budgets, artist concepts and release records.
The part AI does not remove
Nik’s story is useful because it moves the AI music conversation past generation.
Generating a track can be fast.
Building something people recognize, return to and pay attention to is slower.
That requires selection. Direction. Rights awareness. Records. Metadata. Distribution. Positioning. Financial discipline. Patience. And eventually enough evidence to know which parts of the system are actually working.
AI can increase what one creator is capable of producing. It does not eliminate the need to operate what you create.
That may be the more useful lesson behind 80+ releases and 20M+ streams.
The track is not the finish line.
It is where the operating work begins.
