“Music generated by AI will not cannibalize human creations” – L’Express

He is famous among music stars. Less known to the public who listen to these artists. By betting before everyone else on digital, the French Believe has become a global giant in online music, which distributes and produces creators as well known as Jul or Vianney. If it has skillfully surfed the streaming wave, the company is certain: the digital transition of the music industry is only halfway through. Many markets, particularly in Asia, have not yet shown their full potential. And generative artificial intelligence promises to revolutionize musical creation. Interview with the co-founder of Believe, Denis Ladegaillerie and Nicolas Rose, managing partner of the XAnge fund, former historic shareholder of the company.

L’Express: Automatically generated melodies, artist voices that can be cloned at will… With generative AI, will the world of music be shaken up again by technology?

Denis Ladegaillerie : All major platforms, from YouTube to TikTok, are transforming the way they recommend music using AI. The titles are grouped into “clusters” according to various characteristics (tempo, tonality, etc.). The reactions of the audiences too. All of this perfects algorithmic recommendation. However, the major impact of generative AI in music, in our opinion, is that it will be an aid to creativity, a co-pilot. She will analyze large bodies of pieces, develop expertise in what constitutes a quality piece, and advise artists. Like you have Quincy Jones behind your shoulder helping you improve your melody. It is likely that creations with less added value – a background sound for a 10-second ad – will, however, be entirely entrusted to generative AI in the future.

Digital platforms allow everyone to stream music. AI will help anyone who wants to create them. This is exciting, but isn’t the volume of music titles likely to explode? And the artists to find themselves drowned in a frightening cloud of content, invisible so great will be the number of productions, more or less human, more or less automatically generated?

DL: This is a fear of the sector because there will be, it is true, a greater number of contents. But I have a different vision of the subject. Fifteen years ago, the public did not have at their disposal the profusion of titles that they enjoy today. 200 artists in France then captured 90% of the market and the others practically did not exist. Dematerialization has transformed this. The top 200 artists now represent only a third of listenings. At the same time, a fringe of some 7,000 artists has emerged who are now generally able to make a living from their music. For what ? Because the platforms’ recommendation algorithms allow them to find very targeted audiences. We have moved from a mass market to a much more segmented market, supporting more artists and types of music.

Nicolas Rose: We have forgotten it today, but in 2007, when we invested in Believe, it was the time when Jacques Attali predicted that music would become free. This was not verified. For what ? Because a piece does not reach listeners if, at the same time, someone does not do the work of presenting this music to the right audiences. This is the added value of a record company.

DL: The human, technical and marketing resources that we invest in our artists amount to tens of millions of euros. Tomorrow, Internet users will perhaps use AI to automatically generate millions of songs, but this will be of no use, this content will not be listened to. Making quality music with generative AI is actually expensive. Putting 100 million titles through the deep learning mill costs a fortune in computing power. Let’s also not forget that storing and making titles available also costs money. A label that made the crazy gamble of using generative AI to produce half a million low-end titles and then receive royalties on them would undoubtedly lose more money than it would gain. That music generated by AI cannibalize human creations, I don’t believe it for a second.

Before AI, the music industry had already undergone a radical digital transformation. How has technology profoundly reconfigured this sector?

DL: In Europe, we are only halfway through this transformation cycle, and in other countries around the world, streaming is still in its infancy. The transformation will continue for at least another fifteen years. Who says consumption of digital music says development of digital artists. Jul well embodies this new generation of artists who produce regularly, know their audience very well and intelligently use levers like YouTube.

Believe bet that music would move from the analog era to the digital era, starting in 2005, at a time when Spotify did not yet exist. In practice, what new technological tools have you built to anchor yourself in this new world?

DL: We have three fundamental building blocks. The first is content management. Being able to collect their music and videos from our artists, encode them in the appropriate formats and send them to all platforms – Apple Music, Spotify, YouTube etc. Our roots in the hip-hop world stimulated us, because it’s a scene where artists often make spontaneous creations. They go into the studio to make a track or a featuring which does not fit into the defined framework of an album or EP release. And if they leave the studio on a Friday at 8 p.m., they expect us to have their song available on all platforms at midnight. It was a challenge! We have built the right technologies to meet these demands.

The second brick of Believe is the collection of income, much more complex today. Previously, when you sold CDs in France or Europe, you had 4 or 5 sources of income. Now, with streaming platforms, we collect billions of micro revenues around the world, in yuan, in euros, in Indian rupees. We need to know how to aggregate this and give artists understandable tools for monitoring their income. The last brick is our audience development tools. There is of course always a part of human work, of exchanges with the distribution platforms. But technology now plays a vital role. Our tools allow you to see what action to take at what time to have the optimal impact on the audience. Which title to push to the algorithm at what precise moment, when to enter it on a playlist and when to remove it, when to upload a short TikTok video or buy online advertising space. A myriad of technical actions that can significantly improve the visibility of a music title.

What are the most important global markets for Believe?

DL: Today, we are the market leader in Asia and Europe. Within the EU, we are number one in France for local artists and number three in Germany. Our goal is to consolidate our positions in these markets and begin to lay foundations in the United States.

Why not attack the American market earlier?

DL: The American market is the largest market in the world but also the most expensive to develop. It was more strategic to start with Europe, our native region, and with the Asian markets which will become the largest global markets in the years to come. Developing the United States now that we have a critical size and multi-country expertise will cost us less today than if we had started there from the start.

NR: Believe has always had this ability to project itself internationally. Build a centralized platform that allows economies of scale while adapting to local particularities because music remains a highly cultural product. The company also has the originality of growing through acquisition. This is more common these days in Europe because the availability of capital is now better there. But at the time, Believe was a pioneer.

On May 30, the AMF authorized the takeover offer for Believe by the consortium you formed, Denis Ladegaillerie, with the EQT and TCV funds. Why did this buyout seem necessary to you?

DL: The digital transformation of the music sector will continue for fifteen years. And in this context, Believe aims to play a consolidator role, by carrying out significant and transformative acquisitions. The objective was to position the company with a new round of shareholders able to support it during the next five years of growth of the group.

Believe was listed on the stock exchange at a price of 19.50 euros per share. Despite the company’s good performance, its price then stagnated for a long time around 10 euros. A worrying signal for French tech companies dreaming of an IPO?

DL: I think the volumes of capital available in the market are barely around 10% of what they should be. The government understands the problem and is working on it. The question of financing the first cycles of tech companies was well addressed. In the following cycles, there is indeed still work to be done

NR: This trend is not a recent phenomenon. The question now is that of the rise of European stock exchanges and the unification of capital markets. Today we have a good layer of start-ups that have reached a critical mass and who deserve more funding.