The gap between building a demo and building something that scales to 600,000 locations | Mood Media CPO

Trey Courtney started his career as a software developer at Accenture and spent 16 years evolving into the role of Global Chief Product and Partnership Officer at Mood Media, the largest provider of audio and visual in-store experiences in the world. If you have heard music playing in a Macy’s, seen a menu board at Whole Foods, or walked through a Ferrari showroom, you have experienced what Mood Media has scaled to 600,000 locations worldwide. Trey was selected as a finalist for Products That Count’s 2026 CPO Awards, and in this episode of Product Talk, host and Products That Count Resident CPO Renee Niemi sits down with him to talk about what it takes to bring a product mindset into a sales-led, private equity-backed company, how he is navigating AI investment when the stakes are high and the tools keep changing, and what he thinks the CPTO org chart looks like in three years.

Join us for new conversations with leading product executives every week. Roll through the highlights of this week’s event below, then head on over to our Events page to see which product leaders will be joining us next week.

Show Notes:

1. Trey started his career as a developer at Accenture, worked his way into running development teams, then spent time at a joint venture between eBay and Accenture before finding his way to Mood Media. He has been at Mood in its various forms for 16 years, during which the company has been public, private, and PE-backed. The reason he left pure development was not a loss of love for the craft; he still owns engineering today. He wanted to understand both sides of the equation: the technology and how it affects the business, positively and negatively.
2. Mood Media is a company most people have experienced without knowing it. They provide audio and visual in-store experiences for retailers, restaurants, and hospitality brands in about 600,000 locations worldwide. If you have heard music in a Macy’s, seen a menu board in the seafood section at Whole Foods, or walked through a Ferrari showroom, you have seen Mood’s work. Trey describes the company as the brand behind the brand: their job is not to tell Mood’s story but to give brands the tools to tell their own.
3. Mood’s core platform is called Harmony, and it is both hardware and software. But how Harmony manifests at Ferrari versus how it manifests at Whole Foods can look completely different, because the experiences are largely bespoke to each brand. That means Trey’s team has to be deeply close to each client, understand what the brand is trying to accomplish at the customer level, and then design toward those goals. It is more service company than software company in many ways, and he is fine with that description.
4. Mood is a sales-led company, and Trey is direct about that. A combination of factors drove it there: audio has been around since the 1930s and has the characteristics of a commodity product, which tends to push decision-making toward revenue; and private equity ownership creates constant pressure around what returns look like. When Trey arrived, there was no concept of product in the company. As client needs shifted toward integrated experiences, he introduced a product mindset and built the platform infrastructure to support it.
5. The roadmap at Mood does not work the way roadmaps work at most software companies. Trey does not sit down in January and map out Q1 through Q4. A significant portion of his roadmap is consumed by custom requests from large enterprise clients, who are not paying for development directly but still expect features built to their specific needs. That requires his team to spend more time with clients than is typical for product organizations, and it requires a different kind of prioritization discipline.
6. His role has expanded over time from a technical product role to one that includes creative, partnerships, and a broader mandate. Creative is included because content and technology have converged at Mood. Music channels are built to match brand identity, which requires both artistic judgment and technological infrastructure. Having creative siloed from product led to situations where the company sold clients experiences it could not actually deliver, or discovered late that the hardware recommendation was wrong. Combining the functions eliminated those gaps.
7. The AI bet Trey is most excited about is real-time personalization at the in-store level: experiences that use all available data about a retailer and their customers to create something tailored in the moment. He started investing in this direction three to four years ago, beginning with 11 Labs for voice synthesis. At the time the output was detectable as AI and not great, but he wanted to invest early because a significant portion of Mood’s business involves custom audio messaging for clients, a process that currently takes three to four days and involves script writing, client approval, voice talent, and multiple rounds of back and forth. AI compresses that to near-instant.
8. The hardest part of his job right now is not understanding what AI can do. It is deciding where to commit. He could spend all day reading about new tools and approaches. The real challenge is picking a platform to build on and accepting that in three months something better might exist, while also knowing that waiting is worse than committing. His current approach is to move fast on pilots, bring clients into the experimentation, run tests in five or ten stores, and then evaluate before committing to a larger build.
9. Large company AI adoption is not the same as startup AI adoption, and he makes this distinction explicitly. At Mood’s scale, security is a real constraint. You cannot pick a tool, launch it, and hope for the best. Every AI initiative has to go through compliance and security review, which slows some things down while other things move fast. He spends most of his time navigating that tension: where to get quick wins, and where to invest long for the future.
10. He runs an internal AI super user group that includes employees from across the company, selected either because they have domain knowledge in areas where AI could help or because they are technically curious. The purpose is to get people using AI tools to improve their own work, not just to support the product. His philosophy: you are going to get it wrong, there will be experiments that fail, but the only way to really get it wrong is to wait for perfect. Perfect is the enemy of good enough, and the learning window for AI is now.
11. Creative teams at Mood have surprised him by embracing AI more readily than he expected. The common assumption is that creative people will resist AI because it threatens their livelihood. He has found the opposite: creatives are excited because AI handles the low-value internal tasks so they can spend more time on the client-facing work they actually care about. Engineering teams have been more skeptical. Senior engineers in particular push back that prompting Claude repeatedly is not always faster than just writing the code themselves.
12. The place where AI is generating the most excitement at Mood is product management, specifically prototyping. A process that used to take hours of mocking up features to get a client to agree on something now takes close to an hour. He can go from “tell me what you want” to showing concepts within an hour, which significantly compresses the time to get a feature into development. For a company where a lot of the work is custom client-driven, that acceleration matters a lot.
13. A new and unexpected challenge has arrived: salespeople are starting to use AI tools to prototype things themselves and show those prototypes to clients. The problem is the gap between what those tools can spin up for a demo and what it actually takes to bolt something onto a platform serving 600,000 locations with all the compliance, security, and integration requirements that implies. Clients see a demo in a day and ask why the real build takes longer. Managing that gap is a growing part of his job.
14. His answer to the sales and product tension in an AI world is to shift the conversation from features to outcomes. Salespeople have always brought in feature requests; what is different now is that they are also bringing in prototypes and creating expectation gaps. The way out is to stop talking about what to build and start talking about what the client is trying to accomplish. When the conversation moves to outcomes, the complexity of the underlying build becomes easier to explain and easier for the client to understand.
15. Looking three years ahead at the CPTO org chart, Trey’s clearest prediction is that QA and DevOps will change fundamentally. The tools that support testing and deployment automation are going to get so powerful that the timelines and headcount required will compress significantly. He also sees the ratio of product managers to engineers moving closer together, as engineers become capable of more and as AI handles more of what was previously purely engineering work.
16. Agents are not yet on his org chart, but he thinks about them. Mood is building agents that allow salespeople to ask questions and generate recommendations for customers, with the eventual goal of giving some version of that capability directly to clients. The harder questions he is still working through are about orchestration: do you want one super agent or five specialized ones, how do you make sure people use them appropriately, and how do you govern that at scale.
17. The competitive intelligence problem was one he solved with an agent. Mood competes in multiple markets, the competitors vary by geography and by vertical, and salespeople constantly need battle cards for specific competitor matchups. He does not have the headcount to staff that function. So he built an agent, gave it the competitor list and the evaluation framework he wanted, and it now generates battle cards automatically. The output is about 95% accurate and goes to marketing for distribution before being shared with sales.
18. His comparison for where AI is headed is the dot-com bubble. Early in his career, building a website cost millions of dollars. Today you can build one in a day with Wix before you even get to AI tools. The compression of time and cost is real and it is going to continue. The people waiting for the technology to mature before investing are going to miss the learning window, and that window is what gives you the institutional knowledge to use the tools well when they do mature.
19. The book he is currently reading is Careless People, an account of Facebook’s early years of navigating government policy and elections written by someone who ran public policy there. What he is taking from it is a lesson about culture: the difference between the culture a company says it wants and what is actually happening. Facebook was a U.S. tech company that did not want to get dragged into government policy, and the gap between that stated identity and the reality created significant culture problems. He thinks about that in the context of how you build and maintain culture through hard moments.
20. His summary of the AI leadership challenge is the jagged line. If you are investing in AI, you are going to break things. The path to progress is not straight. But as long as you have a north star, as long as you are getting people around you invested in the journey, and as long as you are not sitting back waiting for perfect, you are moving up and to the right. The people who are going to get it completely wrong are the ones who are waiting.