How working producers actually use AI (it's not what the headlines say)

How do producers actually use AI? Mostly for the jobs nobody became a producer to do. The best survey data we have says the most common uses are splitting songs into stems, mastering and EQ, with fully AI-generated songs a small minority. Headlines about AI swallowing music focus on one-prompt song generators; the people who make records for a living mostly use AI as an assistant inside a workflow they still run.

This page is part of the DAW 101 series; the hub is what is a DAW.

What the surveys say

Sources: Music Business Worldwide on Tracklib 2024, MusicTech on Tracklib 2025, LANDR.

Why the numbers disagree so much

32% and 87% sound like different planets. They are measuring different things in different crowds.

• The question. LANDR counts AI used "somewhere in their workflow", which across its 52 tested uses includes cover art, audience research and promotion. Tracklib asks about using AI in making music.
• The crowd. LANDR surveyed music makers in its own global community, the users of a company that sells AI mastering, from beginners to experts. Tracklib surveyed producers, a group that, per its 2024 survey, was 10% professional and 58% "ambitious" about a production career.
• The timing. Adoption is rising. Tracklib went from 25% to 32% in a year, and 69% of LANDR's respondents said they use more AI tools than the year before (LANDR).

Both agree on the shape, though: AI is mostly assistive, and whole-song generation is the minority case. That is our reading across the three, not a figure any one of them states.

The five jobs AI actually does in a producer's day

1. Splitting stems. The top use in Tracklib's 2024 data by a wide margin. A stem splitter turns a finished song into separate vocals, drums, bass and the rest, for remixes, samples, practice, or studying how a reference track is built. It is now built into Logic (Stem Splitter), FL Studio (Producer Edition and up) and Ableton Live 12 Suite (Apple, Image-Line, Ableton). Our what are stems guide explains the format.

2. Mastering and EQ. The second-biggest use. Tools such as iZotope Ozone's Master Assistant create "a tailor-made preset for your audio" that you then adjust (iZotope), and Logic ships a Mastering Assistant. Producers use these as a fast first draft, not a final verdict; see how AI mastering works.

3. Session players and part generators. Logic's Session Players, which Apple calls "AI-powered musicians", play drums, bass, keys and synth parts that follow your chords. FL Studio's Chord Generator and Loop Starter do a similar job for starting points. These are the "generate elements" users in Tracklib's data: around one in five AI users in 2024.

4. Assistants that answer questions. FL Studio's Gopher is "FL Studio's AI assistant that answers questions, guides you, and controls features in FL Studio and plugins" (Image-Line). For beginners, this replaces a lot of tutorial-hunting.

5. A producer in the session. AI DAWs go one step further and put a producer at the centre of the studio. Veena's CoProducer builds beats, chords, melody and arrangement as editable tracks from a plain-English request, and the free plan adds a 4-stem splitter, exports with no watermark and real-time collaboration (Veena). Used well, it is the session-player idea taken to a whole song, with every part left open for the producer to change. Our working with an AI co-producer page shows a session step by step.

What producers will not hand over

The same surveys show the limits producers set. In Tracklib's 2025 study, 82% opposed creating full songs from text prompts, 78% said fully AI-generated music should not be on streaming services, and 83% supported mandatory disclosure of AI use (MusicTech). In the 2024 survey, 82.2% of non-users gave artistic reasons, such as "I want my art to be my own" (Music Business Worldwide).

That is not anti-AI. It is a clear boundary: help with the work, but the song is mine.

What it means for a beginner

• Copy the pros. Use AI first where they do: splitting references into stems to study, getting a draft master, cleaning up a noisy take.
• Use generation as a teacher. If a producer in the session writes a bassline, open it and change it.
• Keep it editable. Tools that give you tracks and MIDI keep you in charge. A finished, locked render does not; see AI DAW vs AI song generator.

What we think

The headline story, AI replacing producers with a prompt box, is the least common use in the data. The real story is quieter and more useful: AI took over the tedious parts of the job, such as separation, cleanup and a first-pass master, and is now moving into the studio as a collaborator that drafts parts a producer edits. We think that second step is where the most value for new producers lies, because it turns every draft into something to learn from. The surveys also tell you how to stay on the right side of the craft's own line: let AI assist, keep the decisions, and be open about what you used.

Previous in the series: is using AI to make music cheating? Next: collaborating in a DAW. Related: will AI make producers lazy?, our DAW AI features tracker and AI in music tech beyond the DAW.