As a drummer, I recently needed to replace some parts on my Roland digital drum kit. Little did I expect that doing so would open a series of conversations with the C-suite of what is probably the least well-known e-commerce innovator — and the largest music retailer — in the US.
I placed my order online at Sweetwater.com on May 6th. My order shipped out 24 hours later, and it was accompanied by an email from my so-called “sales engineer,” Justin Poole. Thinking this was maybe a mistake, I didn’t follow up. Shortly thereafter, I received a voicemail from Justin letting me know the shipment was on its way and should I need anything else in the future, he was my guy. That’s when I realized there was something happening here. First-class person-led customer service in a time when the typical enterprise ambition for all pre- and post-sale services is to become 100 percent automated and agentic.
My bandmates had had similar experiences; my bass player used the term “stalked” when he talked about how attentive his sales engineer was, and he promptly sent me a few funny memes.
When my new snare and bass drum arrived, my experience confirmed something else my bandmates had mentioned: Sweetwater products always ship with a little complimentary candy. And the unboxing sweetness continued when I got another email from Justin, making sure everything looked right.
The Origin Story
Today, Sweetwater is a $1.9 billion annual revenue company with 600 sales engineers and 13 million customers. The physical headquarters outside Fort Wayne hosts a one-million-square-foot facility that supports recording studios, classrooms, and multiple sound stages for live events. The company was named one of the Global Top 100 Inspiring Workplaces 2025, and gives back over $1 million in charitable donations each year.
The seminal idea for what would become Sweetwater began in the mid-1970s when an innovative touring musician named Chuck Surak built a recording studio out of his Volkswagen bus. Surak was recording bands in the Fort Wayne and Midwest region. By the early 1980s, in one of those “right place, right time” kind of occurrences, Surak got to know Ray Kurzweil, the inventor of one of the era’s breakthrough music technologies: a digital sound sampler called the Kurzweil K250. The K250 was becoming known as the next big thing in the music industry. There was even a braille version made for Stevie Wonder. It was the first digital instrument to really sound like an acoustic piano. Surak got to know Stevie Wonder and others who recognized him as “that whiz kid from Fort Wayne,” someone pushing the boundaries of music technology.
As both drum machines and sound samplers began to expand the possibilities for musicians and recording studios, Surak saw the business opportunity, and began to rep companies like Kurzweil — both selling the instruments as well as providing technical support services.
This weaving of retail and service is Sweetwater’s DNA. To better understand how it’s evolving in the AI era, we recently spoke with Sweetwater CEO Mike Clem and CIO Jason Johnson.
• • •
Sweetwater’s CEO on the Importance of Customer Connectivity
Mike, you were Chief Digital Officer, and then became CEO in 2023. You are continuing quite a legacy. Clearly innovation runs through the company. How is it shepherded today, fostered and inculcated around the firm?
We believe innovation is just baked in culturally to what we do. First of all, sometimes people use the word innovation, [and] they think it’s just a technology play. No. We would consider innovation in any department, any function. And we have built a culture that rewards that kind of thinking and we hire those kinds of people. We describe ourselves as a culture of “builders.” We certainly partner with outside folks, but whenever we can build ourselves, that’s the natural thinking here. We think in systems and we think in a framework we call 60-30-10, where 60 percent of our energy at any moment should be focused on here and now, things that produce immediate value. We reserve about 30 percent of our energy for innovation that might mature within, say, the next three years, and then we reserve 10 percent of our energy for what you might consider R&D, or just kind of the long arc.
Let’s talk about the specifics of innovation and foresight, getting ahead of what current customers say they want and exploring what they might need next. Who is responsible for product and new business model ideas?
We have an R&D group working with partners on some next-gen music-tech initiatives. …But [it’s] no question [that] our special advantage is that we’re on the phone with our customers, truly listening, hearing. We are closer to customers than most retailers would dream to be. And we are sitting on about 45 years’ worth of this kind of rich data. And so we get to turn that into insights to help drive our innovation roadmap in a more sincere way than a lot of retailers get to do.
We’re in a high-passion industry, where customers don’t want to talk to robots about guitars and rock and roll. So much of what we do is personalized advice, compatibility, complexity.
CEO Mike Clem
Speaking of 45 years of rich data, where is AI in your company strategy, and how do you see it shaping or changing how you work with customers going forward?
We’re in a high-passion industry, where customers don’t want to talk to robots about guitars and rock and roll. So much of what we do is personalized advice, compatibility, complexity. AI is not quite there, [in terms of being] able to understand needs, and provide that level of expertise and consultation that we do. So will it replace call centers for other kinds of retailers? Probably so. Will it replace the kind of expertise we provide? I don’t think so.
Outside of your musician customer base, you’re focusing on digital creators. You now sell drones and other gear for them.
We are seeing just unbelievable exciting growth in podcasting, gaming, streaming, influencers, etc., serving the video and camera side of that. It’s evolving and exploding open very quickly. You think about the pandemic, when live music went to zero. [People] turned to streaming, [or] went to YouTube. Now that live music is back, many musicians continue to maintain those channels. And so now we’re all creators in that sense, and there’s a whole new ecosystem opening up.
Would you describe your approach as open innovation, where you bring outside partners together and explore the creation of new products or even categories?
Not formally as such, but we put our physical campus to good use. We have a music academy, [and] a world-class recording studio with major artists here all the time. We have an amphitheater where we’re doing concerts throughout the summer. We have artist workshops. We want to create an environment where the artists can come in, the brands can come in, we can collaborate. Sometimes we have 10,000 to 20,000 customers here at a time for big events. We just believe it’s important for the industry [to] create that community, and not just push boxes out the door.
..We have partnerships with all the big technology companies, especially around marketing tech. I would say we have a lot of partnerships like that where we’re developing the next generation of retail technology or marketing technology integrated into building what we call our customer data platform, where we keep those 45 years of customer understanding, pouring AI and data science over it.
We have a lot of partnerships [with] music schools [and] early-stage music tech startups, too, where we’re doing some really fascinating things with companies working on immersive sound, [and] one partner company doing fascinating things with virtual reality capturing sound differently, reproducing it through transducers instead of speakers. We love partnering with these young music tech companies, and just bringing innovation to the industry.
Sweetwater’s CIO on AI Strategy
Jason, as CIO, give me some top-line numbers from your side of the house?
We deploy code about 400 times a week to production. We are very much an iterative shop, constantly improving things. …The team is about 240 engineers. …It would not be uncommon if you toured our distribution center right now to see one of my developers sitting at a pack station or a picking station, programming with AI in real-time. I really feel very strongly that when you write proprietary software for the business, you have to get out in the business and really partner with your customer.
Your CEO describes Sweetwater as a place of many creative experiments and external partnerships — and a place with incredible first-party data. How are you leveraging that data to build your own in-house AI?
We are just super, super blessed to have all these human beings that have generated all this data over 40 years. I truly believe in my core that AI is a tool, not a replacement for humans. Not in our industry. I think human expertise and human creativity are going to become more and more important as we move forward.
Our in-house AI is called Echo, and it’s our proprietary chatbot based on Claude. Here’s an example of how it’s getting used: we have a group that does a particular leadership course run by our HR department [over] 13 weeks. Through every course and every class, they do surveys. At the end, there’s somebody who usually spends hours aggregating all of that into useful reporting for the executive team. And that person wrote me a wonderful note a couple of weeks ago: “Hey, I did this for the first time and I used AI. I dumped all the surveys in there, and it built me a beautiful report in minutes that had more actionable data in it.”
The AI startup Glean recently did a survey that noted an average of 11 hours of work a week saved for many knowledge workers, but much of the saved time was spent babysitting those AI bots. How do you make the productivity and time saved story real?
The biggest question I ask back to the business is, “OK, but what do we do with the other 3.5 hours that you have free now?” Part of you just hopes that all that time is getting dumped right back into the business. I think AI is as fundamental as the tractor. And when farmers got tractors, they didn’t just plow their three acres and take a nap at 9 A.M., they went and got 200 acres. So the question for me is, how do we create these loops where we can recognize…that some work is moving faster, and we can get employees and teams to reinvest that horsepower back into the engine that will amplify it and allow us to serve more customers.
How are you thinking about AI strategy — the agentic tools, and how it gets applied to your growth and operations — without freaking everybody out?
We have two different AI groups going.
In January, I built a small group that works directly reporting to me. We call it the Emerging Business Technology Team. They’re mostly doing AI. It’s like our little internal applied AI group. It’s just two people, and they’re moving mountains. A lot of their work has been injecting themselves into workflows and processes inside Sweetwater and asking the question, “How can AI accelerate it?” I’m looking into the existing metrics that have run our business for 44 years, and we’re going to make those metrics go bigger and faster and stronger.
Then, we have another couple of folks that are focused on…how do we arm people with how to use [these new tools], and use them responsibly. And frankly, AI has become a political-level conversation that we typically try to avoid at work at all costs. No one talks about Microsoft Word like that. No one talks about Excel like that. So how do we tackle that, and at the same time train people, find people that are doing good work, and amplify that?
So what’s your solution to both allaying fears of AI, and circulating or sharing breakthrough uses?
I think the solution is one-on-one training, group training, information, and knowledge. Tackling those questions head-on, and arming people to go do their best work. I’m not sure any company is going to rule the world by just putting ChatGPT or Gemini in everybody’s hands, and saying, “Go forth and conquer.” The question becomes, how do we get people to embrace and adopt the roles they’re good at, and talk about the things they’re not good at? That all starts with asking, “What is the toil happening in your day-to-day job that, if we were able to remove it, could it allow you to focus on your most important work?” For many people, specifically our sales engineers, that means more time for relationship-building and writing better solutions.
What do you count as your biggest wins to date, where the needle is moving in the right direction?
We’ve accelerated software development per team somewhere in the range of 20 to 30 percent. We’ve added almost 30 virtual full-time employees to our software development teams, and we’re seeing that work happen. That 20 to 30 percent is lumpy. There’s one gentleman who did almost 60 days’ worth of estimated development work in like five days, and a lot more people doing 10 days in their seven days.
There is of course an expense to that — not just the hard token [AI usage] costs, but also we’re finding business units that we serve that are going, “Hey, you’re actually making so much change in our department so fast now that we’re not ready to test it all.”
But overall, it’s been super-exciting.
How else do you close the loop on all this AI learning?
One of the things we did was create a program called Token Exchange. Our question was, how do we get somebody who’s really good, [and team them] up with someone else? Not through a lens of, let’s sit down and talk about it, but let me sit next to your desk and work with you for a little bit and work through some tasks. Maybe you figure out that AI is not good for the tasks you’re working on. Maybe you go, oh my goodness, this is a gold mine.
Also, a few things that we haven’t launched yet but are in the pipeline: an enterprise skills repository. And for a company our size, you have to create an internal marketing play. I’ve got to communicate this out with everyone to help them change. And I do think part of it is just talking about it all the time and keeping it top of mind. I’ve seen people’s desks with a sticky note on their monitor that says “Can Echo help?” If you’re just coming in every day and doing your thing, it’s really easy to not try.
I have a personal Claude Code agent. He has his own computer. …I think that is the future.
CIO Jason Johnson
You have RAGs and vector databases doing inference over tons of content. But what about good old-fashioned machine learning (ML), i.e., traditional AI? Is that still in the mix?
Yeah, for sure. We’ve been doing machine learning for over 10 years. We’re doing that at an accelerated rate now. Generative AI can actually help you accelerate machine learning. So we’re all in on that, for things like product recommendations and fraud models. And we have a lot of internal machine learning use cases, like how we match customers with each other. We don’t call it eHarmony, but that’s how I kind of think about it, how we match sales engineers and customers.
Outside of work proper, how are you personally staying on top of all this, when it’s changing so fast every day?
I have a personal Claude Code agent. He has his own computer. A couple of people at work are also doing it. I think that is the future. Having autonomous agents that are very skilled sitting in the corner, kind of waiting for work and mining work, that’s definitely a successful model. I think you can do that with some of the off-the-shelf tools.
My Claude Code agent has his own ID through [the cloud communications platform] Twilio. So I can call my agent and can talk to it through that channel. I think that just like having an army of real people — living human beings working for you — you have to think about these as additive things. Managing a bot is actually work itself, and probably the future of work for a lot of people in engineering roles, [if] not eventually everybody.
You must be logged in to post a comment.