For Headspace, design and research aren’t support functions. They’re the connective tissue between member needs, business decisions, and long-term impact.

In a conversation with Michael Winnick, CEO of Dscout, members of Headspace’s design and research leadership team—Cal Thompson, VP of Design; Jonathan DeFaveri, Design Director; and Chelsea Coe, Head of Research, Product & Design—shared how research has shaped some of the company’s most meaningful product decisions.

From deeply human member stories to AI-powered workflows, the conversation offers a clear picture of what happens when research is embedded into how teams actually work.

Research that stays grounded in real lives

The conversation opened not with metrics, but with moments.

Each leader shared a story about hearing directly from Headspace members—stories that continue to shape how they see their work. A man navigating grief after losing his wife. A former classmate relying on Headspace during a hospital stay. A friend finding peace after miscarriages through daily use of the app.

These weren’t curated testimonials or polished case studies. They came through always-on feedback, personal conversations, and research sessions. And that’s the point.

“When you hear these stories,” the team reflected, “you don’t forget them.” They become reference points, anchors that keep product decisions grounded in empathy, not assumptions.

Designing a refuge, not just a product

Headspace has always had a strong design identity. But as the company evolved from meditation content toward a broader mental health platform that includes therapy and AI-driven support, the team faced a new challenge: how to evolve without losing what made Headspace feel safe and approachable in the first place.

For Cal, that meant returning to a core question: How can the app feel like a refuge?

Design choices, from color and illustration to voice and interaction patterns, are all guided by that idea. The goal isn’t to maximize engagement for its own sake, but to respect members’ attention and emotional state—especially during moments of high need.

That principle became even more critical as Headspace expanded into higher-acuity care.

The research insight that changed everything

One of the most pivotal moments discussed came from a Dscout diary study run during the height of the pandemic. Over 30 days, the team followed members’ daily experiences to understand how Headspace fit into their lives.

One participant described the app as “that boyfriend who says he’ll always be there for you and then disappears when you need him most.” The quote stopped the team in their tracks.

It wasn’t a usability issue or a feature gap. It was a signal that some members were coming to Headspace in moments of distress that meditation alone couldn’t support.

That insight helped catalyze a major business decision: Headspace needed to expand beyond content and into therapy and broader mental health care. Research didn’t just validate the direction—it made the need undeniable.

From big studies to everyday research rituals

As Headspace scaled and designers became more embedded within product teams, research had to evolve, too. According to Jonathan, the team no longer had the luxury of long, stand-alone studies, but skipping research entirely wasn’t an option.

The solution was Koi Pond: an internal, standing, monthly research ritual built on Dscout.

Koi Pond functions as a trusted member panel where teams can quickly test early concepts before investing in higher-fidelity design or development. Recruitment, screeners, and workflows are already in place, dramatically reducing setup time.

What changed wasn’t just speed; it was behavior. Designers, product managers, and engineers now ask, “Should we run this through Koi Pond?” Research became a shared checkpoint rather than a specialized request, embedding insight directly into product momentum.

Making synthesis faster and less intimidating with Dscout AI

That momentum matters most when time is tight.

Chelsea described moments when deadlines were measured in hours rather than weeks.

“We were on the heels of a very fast deadline—like hours, basically due the next day—and I didn’t think we were going to be able to do it. Using AI was really the only way we could have made it happen. It helped us move past that blank-page fear and actually get started.”

Dscout’s AI features didn’t replace the team’s judgment or craft. Instead, they created traction at the moment it mattered most.

“It wasn’t about copying and pasting,” Chelsea explained. “AI showed me what everyone was saying in aggregate, and then let me go deeper into a single topic and pull the exact quotes we always include in our presentations. That really streamlined the whole process.”

For Chelsea, the value extended beyond speed. AI lowered the emotional and cognitive barrier to synthesis, especially for non-researchers.

“I can see this being incredibly useful for people who aren’t researchers. A lot of designers are really fearful of the synthesis process, and this makes it feel less scary—like you have help already embedded in the work.”

Seeing an entire study come together in one place changed how the team experienced analysis itself.

“I loved seeing our entire study light up in AI, summarized step by step. It was really cool to watch the whole thing come together in a way that felt fast but still meaningful.”

The work stayed human. The heavy lifting got lighter.

Building trust with empathetic AI

That same research rigor carried over into Headspace’s own AI product: Ebb, an empathetic AI companion inside the app.

From the start, the team used research to answer a critical question: not just where AI could help, but where members trusted Headspace to use it.

Ebb was designed to feel warm, credible, and distinctly Headspace. Research informed everything from tone and interaction style to how recommendations surfaced. The result was an experience that felt like a private corner of the Headspace universe, one that was intimate, safe, and supportive.

Research as culture, not just practice

One of the most telling moments came when the team described “member week,” a series of live interviews recruited through Dscout and broadcast internally across the company.

For many employees, it was the first time they’d seen members speak openly about their experiences. The effect was immediate. Teams across marketing, product, and leadership began advocating for more. At Headspace, research didn’t just inform decisions. It aligned the organization around why the work mattered.

Growing together

As the conversation closed, Michael Winnick asked what Dscout should never lose as it evolves. The answer wasn’t about features; it was about supporting researchers as creative partners. That includes making qualitative research more accessible, less intimidating, and easier to act on, especially as AI reshapes how teams work.

For Headspace, Dscout isn’t just a platform. It’s infrastructure for empathy at scale. For teams designing in complex, deeply human spaces, that makes all the difference.