Tag: biking

The generative artificial intelligence boom has been so rapid and so widespread that you could be forgiven for feeling like the technology has been around for much longer than it actually has. Indeed, November 30 will mark only the third anniversary of ChatGPT’s launch—a watershed moment that kicked off not just the consumer chatbot craze but a much wider effort, across the global economy, to weave AI into nearly every facet of business and commerce.

Still, it’s early innings for the software, and many businesses are still figuring out what (if anything) it means for them. One such company is Zwift, the e-biking and virtual fitness company, which is now in the process of incorporating AI-driven personalized content recommendations into its consumer products. In a recent conversation with Inc., CEO Eric Min noted that the company is really “just one year into real AI in terms of how [they’re] delivering that to customers.”

“We’ve been using it internally for engineering for a bit longer,” he added, “but we’re pretty excited about how this can change and enhance the experience for the customers going forward.”

The chief executive spoke further with Inc. about his thoughts on AI—including where it fits into his company’s post-layoffs rebound and what it means for the broader labor market—earlier this month. Below is a condensed version of that conversation.

In February 2024, Zwift had layoffs and your co-CEO left. You said last fall that you were looking to scale back up again in the wake of that. What has the last year looked like for you in terms of scale?

The last 18 months, the company’s been really performing. It’s the beauty, sometimes, of operating a smaller team and having fewer layers of management and staying really, really focused. We basically said no to lots of different initiatives and focused on just a few things that we thought were material—and that’s starting to pay dividends now.

Can you give me examples of stuff over those 18 months you’ve said no to?

We’ve been toying around with rowing, for example; we pulled the plug on that. We said, ‘We’ve got more important things to do.’ So that’s been shelved; might be shelved forever. Another example is, we really scaled back on running, which we’ve had for quite a number of years. It’s still there, but it’s not a paid service. Our focus really is just our core audience: people who just want to ride their bikes. There was some work that we wanted to invest in around personalization. There’s a big theme around, ‘Tell me what to do next.’ Consumers just want to be told. And there is so much to do in Zwift; that is both the curse and one of the strong points that we have. We have just a ton of content. So the way Netflix and other streaming services provide you [recommendations], or Spotify comes up with playlists for you, we’re trying to do that using AI. So we’re making a big investment there, and that will start rolling out this year.

Aside from the product applications of AI for content recommendation, do you guys also use AI internally?

We’ve been using [Microsoft] Copilot for some time now; our engineers have been taking advantage of that. More recently, we got a corporate license for ChatGPT, for example. We also have Google Gemini. We want our employees to take advantage of these corporate AI tools that are available. It’s just so efficient. There is so much more we can do; leave out all the mundane work, and we want to focus more on, like: ‘What does a customer want? What’s a great design?’ It’s kind of frightening how fast these tools are evolving, and you can do so much more with less staff. It does create some issues around staffing. I think this is true for many industries: I think it’s just getting more and more challenging for graduates. Where do they slot in when you need fewer people? I think this is something that we need to figure out, and I think the industry [does] as well. We just need fewer people to do way more now.

How are you thinking about hiring and headcount in the context of increased AI capabilities?

We’re definitely hiring in the AI space; that’s one area. But what we’re finding is AI is allowing us to operate support, for example, way more efficiently, at scale. So that’s just coming down. And also quality tests, automation—we just don’t need as many people. This is the case for lots of businesses, so I’m excited, but I’m also, on the other hand, a little bit concerned about how the whole labor market is going to shift as a result.

Have you done anything on the content generation front for the biking courses or for world-building?

We’re playing with some of those tools; we’re not there yet. One of our strengths is creating really interesting virtual worlds, and I don’t think the tools like Sora and others out there are just there yet. It’s coming; I still think we need game designers to come up with something really creative. And what you could do is use tools to help aid in their development of art assets. But I think ultimately you still need people to come up with great, great designs.

The team did a fabulous job, and it takes a lot of creative minds to come up with that. It’s not just, ‘Let’s replicate Prospect Park.’ They’ve done really creative ways of connecting, you know, Manhattan to Brooklyn, and I don’t think AI could create that for us. That requires real artists to come up with some great ideas. But we do see a future where these artists that we have—which, frankly, I think we have some world-class artists on our team—they’ll have better tools, and these tools will generate the assets that they do manually today. But I think you still need that creative direction from these artists. So whether it’s artwork, whether it’s coding, I think there are other kinds of content that we can think of that could be generated with AI tools. So we’re just at the beginning.

The final deadline for the 2026 Inc. Regionals Awards is Friday, December 12, at 11:59 p.m. PT. Apply now.