Tell us about yourself. What are you working on right now?

I’m Henry Skelsey, and I lead product marketing for Songtradr. It’s the world’s largest music company for businesses—we help artists, labels, rights holders, brands, agencies, and production studios monetize, and find the music and the partners they need to be successful. For example, our platform, Songtradr Studio, is a web-based tool that lets any brand agency, production studio, you name it, use our AI-powered search to find the perfect music for their content, be it a social media video or a digital ad. 

We also have creative agencies like MassiveMusic and Big Sync Music that work with advertisers to select, compose, or re-record music for their campaigns, for their brands, and for what I find most interesting—their sonic identity. People don't necessarily know what [a sonic identity] is, but the second you hear it, you're like, “Oh yeah!” You know that Coca-Cola has a sound, and Intel has a sound. That sound that plays at the end of every TikTok video? We designed that, and it’s now, I believe, the most-heard sound in human history—played a couple hundred million times a day.

That sound that plays at the end of every TikTok video? We designed that. It’s now, I believe, the most-heard sound in human history.

We cover quite a bit of the music industry at this point. We’ve been very acquisitive over the last decade—most recently acquiring Bandcamp in October 2023, and before that 7Digital, a music delivery platform. We believe that by finding the best talent and technology in the world we can make the music industry a more equitable, fair, and ultimately more creative place for everyone involved.

What’s your own hands-on experience with generative AI?

[My first exposure to] generative AI, like a lot of other people, was the original DALL-E when it came out two or three years ago—it got passed around the different chat groups I was in, people were playing with it, making images and videos, and so I started playing with it too. Then, of course, things went quiet for a bit until ChatGPT hit the scene, and that’s when I started seeing what you could accomplish with it.

I use ChatGPT as a way to get over writer’s block. If I need to write a press release or instructions for our sales teams, or I’m working on some kind of strategic asset, just to get the juices flowing, I’ll type a prompt into ChatGPT to get some output. I don’t use what’s coming out of there, but it certainly helps me get over that initial step, to then go out and do my best work.

Do you use AI in any other part of your product launch process?

Not as of right now. There are some clear, “easy” opportunities for some kind of analytical AI to come in and do mass user research and market analysis, for example, but I haven’t seen those tools yet—or, if I have, they haven’t been very good. 

How does Songtradr use AI?

We bought a German AI company called Musicube around two years ago. Their technology allows us to analyze any composition, breaking it down into individual multi-second segments with thousands of points of data, and then connect it to other external data sources so we can then map how not just that track, but individual sections of track, resonate with—pardon the pun—different segments of the population. We can say, “If you’re trying to engage with women between the ages of 35 and 45 in the Midwestern U.S. with middle-to-upper incomes who are interested in A, B, and C, then these are the kinds of tracks you want to use on social media right now to achieve X.”

However, we are very very thoughtful and meticulous about the training data our AI uses. Our mission is to maximize the value of music for everyone, and this always starts with the creators. So never using unpermitted datasets, and taking a creator first approach is crucial to achieving our mission. 

You’re talking about music in a way that makes it sound like something rational and logical—like a formula—but typically art is perceived as a product of unpredictable inspiration. How do you reconcile those two competing takes?

That’s the great question, right? Not just for music, but for design, for visual arts, for writing—how do we make sure that the beauty of human creativity isn’t crushed by what is, ultimately, a phenomenal data engine? To you, to me, and probably to the average person that rational approach means we lose the joy, the beauty, and everything else that makes the human experience worth living.

I think we have to be very intentional about where we let these tools take over, and I use the word “tool” purposefully. The huge temptation that we’re now seeing with AI is to use it as a replacement for creativity, but when I think about how a creative person might use it as a tool, well, I started my career as a writer—I wrote for a lifestyle magazine in Shanghai—and in the creative process inspiration comes randomly. There’s no rhyme or reason to it. You can’t just sit down and say, “Okay, a new idea, let’s go”—but you can let an AI tool augment that process to some degree. 

How do we make sure that the beauty of human creativity isn’t crushed by what is, ultimately, a phenomenal data engine?

It’s not really that new of a concept—a lot of people have writing partners because the process of bouncing ideas off each other spurs new ideas. Could an AI play that role? For example, we proactively want to empower music supervisors to work better. The process these guys go through is pretty incredible—they watch something, start thinking, “I need a 1940s bluegrass tune that’ll work well here,” and then they spend hours digging through tens of thousands of songs, trying to find what’s in their head. An AI tool could instead be prompted with, “I have this idea, can you find every song in my standard 25,000-song catalog that fits?” That significantly shortens the workflow. They can still do the creative process of digging through the options the AI presents—they can still go crate diving, as record fans say—but they don’t have to spend 45 minutes typing different words into a search engine to get started. They can get right into the actual creativity of their job.

So you see AI as increasing the efficiency of creativity, rather than replacing it?

Yeah, exactly. When I was a writer, I might have worked on an article about some new cuisine arriving on the Shanghai scene, but the research would take hours. If I’d had a tool I could ask, “When did the first Ethiopian restaurant open in Shanghai?” so that I wouldn’t have had to spend hours Googling… My god, that would have made my life so much easier. I could’ve then just focused on the actual writing, instead of all that research.

What would your advice be to people who see AI as competition?

This is a societal-level conversation we need to have. Given the opportunity, the moral hazard we have in a capitalist economy is that people are going to use AI as a replacement for human creativity because it’s cheaper and the output can be greater. Sure, the quality might not be the same, but at some point you’re willing to sacrifice that in the name of lower costs.

Transparency, transparency, transparency is key.

Is that the way we want to go? If not, we need to take steps to control it. If you’re a science fiction fan like I am, you’ll know writers have been exploring the dangers of AI for nearly a century. Asimov wrote I, Robot in the 1940s, and one of the cornerstones of those stories is that humanity banned sentient robots on Earth because of this exact problem. That was 80 years ago—this is not an unknown problem, it’s one that some of the great minds have been thinking about for quite some time. We just have to follow that path.

How do you handle those concerns when trying to market to the musicians you want to convince to sign up and contribute to your AI-powered tools and platforms?

You have to be absolutely, brutally transparent, and you have to give them the option to opt in or out.

What we want to do is open the door for artists if they choose. If you want to take the next step in commercializing your music, then it’s about giving you the opportunity—but not the obligation—to have your music on our bigger, more corporate platforms, and to start working with brands and agencies and studios in a way that our AI-powered search tools can find and discover and recommend.

In every step of that process, transparency, transparency, transparency is key, and at every step it’s optional. It’s never a foregone conclusion. You have to agree to it, and we’re going to tell you exactly what you’re agreeing to.

How do you approach product marketing more generally when it comes to products that integrate AI in some way? 

AI’s good, but it’s not that good, so we want to be very cognizant of its limitations. We need to be honest with our customers: “Hey, this is a tool that’s going to blow your mind and do incredible things in these specific use cases… But otherwise, it’s not going to work so well.” And I know marketers don’t like to talk in such exacting terms because it can give the impression of a limited product, but I’m a fan of being honest with your customers because they’ll reward you for it in the long term. Customers have, thankfully, woken up to what AI really is and what is just marketing fluff, so be honest about what’s an actual AI product and what’s just a million nested “if-then” statements.

It’s kind of like fire. It’s an incredible tool, and it can be both beneficial and destructive.

We have to consciously make choices in how we position and sell our product, and in how we work with our product engineering partners to build products in such a way that they avoid some of the moral hazards we’ve spoken about—to augment people’s work rather than replace it. It’s a conscious decision that you have to make: “I could sell a tool that’s going to put a bunch of people out of work, but I’m going to choose not to.” Some of my more aggressive, more competitive friends would say, “You’re just going to be out-competed in the long term.” I disagree. I think there are always going to be people who prefer to go with human creativity—and [who] are going to seek out the tools to help enhance it rather than erase it. 

Do you have a hot take on generative AI?

It’s kind of like fire. It’s an incredible tool, and it can be both beneficial and destructive—it’s up to us to decide which way it’s gonna go.