Sourabh Pateriya, CEO of Soundverse, Product Leader building ethical and controllable AI music creation platform.
getty
In April 2026, Deezer revealed that nearly 75,000 fully AI-generated tracks were arriving on its platform every day. That figure accounted for 44% of all new uploads. In the same announcement, the streaming service noted that those tracks were getting just 1% to 3% of total streams. The machines were producing the music, but people were mostly not listening.
That single contrast tells you more about the future of generative AI than any keynote demo will.
The dominant narrative says AI is here to replace people. It will write the code, draft the contract, design the deck and compose the song. Investor decks reach for it, headlines amplify it, and road maps get built around it. But when you look at how working professionals actually use these tools, the story inverts.
When it comes to how music producers integrate AI into their work, 60% use AI as an ideation tool, generating melodies, chord progressions and arrangement starters. Thirty percent use it as a co-producer, weaving AI suggestions into final tracks alongside their own playing and programming. Only 5% delegate full production to AI.
Call it the 60/30/5 Rule. The bulk of the value lives in ideation. A meaningful middle lives in collaboration. A thin sliver lives in full delegation. And the producers who use these tools heavily are not the ones generating finished songs from a prompt. They are the ones using stem separation to rescue an old recording, AI-assisted mixing to balance a chaotic session or MIDI generation to break a writer’s block at 3 o’clock in the morning.
In my work at Soundverse AI, I see the same shape every week. Power users rarely press a single button and walk away with a song. They prompt a melody, regenerate it twice, drag it into their DAW, replace the bassline by hand, ask the model for a different drum feel, accept half of what comes back and reject the rest. The work is bidirectional. The AI is the junior collaborator, and the producer is still the artist.
There are three reasons the 60/30/5 distribution holds. The first is judgment. Producers are paid for taste, not output. When a producer rejects an AI-generated chord change, that rejection is the actual work. Then there’s accountability. A producer’s name goes on the final song. They cannot ship something they did not actually shape. The third is identity. The people who chose music as a career did not do it so they could become editors of machine output. They want to be faster, braver versions of themselves.
The pattern is not unique to music. Developers using GitHub Copilot accept inline completions and ignore "build me an app" prompts. Lawyers using tools like Harvey accept clause-level suggestions and do not ship autopiloted contracts. The professional market, across categories, lives in the 60 and the 30.
So, what can builders take away from these trends in the industry? I recommend four practical shifts.
The most important moment in your product is rarely the first output; it is the second, the moment after the user wants to change something. In music, for example, that means stem-level control, region-level regeneration and the ability to lock what is working while iterating on what is not.
Sixty percent of usage lives at the brainstorm layer. That means fast iteration, low cost per generation and easy ways to throw out nine ideas to get to the 10th. Producers do not want a polished song from a prompt. They want 100 half-formed ideas they can sift through to create something that is theirs.
People need to feel that the final work still belongs to them. That means preserving edit history, attribution and version control so users can clearly trace how something evolved. In music, producers need to be able to point to the final track and credibly say they shaped it. The moment the platform appears to claim ownership of the creative process, trust starts to erode.
The story that lands with experts is not that AI does their job. It is that AI removes the parts of their job that are tedious or they don’t enjoy doing, so they can do more of what they are good at.
There is a real market in that 5%, but it’s a fundamentally different market from the professional one. The professional market lives in the slow, unglamorous work of making expert workflows faster, more confident and more recognizably the user’s own. The teams that conflate the two are the ones that risk losing product-market fit. Build for the 90, and the five will take care of itself.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。