Mental health has reached the center of national conversation in U.S. healthcare, driven by a growing crisis: more than one in five Americans experience mental illness each year, yet access to high-quality, timely care remains a daunting challenge. With clinician shortages, mounting costs, and the residual effects of the pandemic, health systems are searching for ways to scale support, identify risk earlier, and move upstream—without overburdening already strained workforces. Enter artificial intelligence.
In this episode of The American Journal of Healthcare Strategy podcast, I sat down with Josh Pappas, Senior Head of Sales at Kintsugi, to explore a technology that’s breaking new ground: voice-based AI tools that detect signs of depression and anxiety from ordinary conversations. The twist? Kintsugi’s approach slots into existing workflows, aiming to supplement—not supplant—clinicians while providing real-time, objective data about patients’ mental health.
Below, we unpack how AI is being woven into mental health care, why career agility matters in digital health, and what executive leaders need to know to ride this next wave of innovation.
AI-driven mental health tools matter now because they tackle two pressing issues at once: a severe shortage of clinicians and a lack of objective, scalable screening methods. As Pappas explains, “Mental health is top of mind for everybody because it's not just, ‘Hey, great, we've cared for everybody's physical health’... if we don't cover mental health, costs are higher. We're not doing a great job now.”
Most U.S. systems still rely on questionnaires or self-reports—the PHQ-9 and GAD-7 are gold standards, but they’re often skipped, rushed, or only used after problems surface.
Stigma, workforce shortages, and siloed data limit early identification and preventive care.
Kintsugi’s innovation is an AI tool that analyzes “subtle changes in how we produce sound and speech”—milliseconds of inflection, pace, and tone—in ordinary conversations. The goal? Surface risk for depression and anxiety even when it’s not explicitly discussed.
“The amazing power of AI [is] looking at subtle changes in how we produce sound and speech and be able to correlate that with depression [and] anxiety… to scale access to starting the conversation around mental health way earlier.”
This isn’t a consumer app or a heavy lift for IT teams. As Pappas puts it, “Kintsugi has built an API so we've built a tool that can kind of embed into what already exists: in the call center, in the virtual care center, embedded into an app... frictionless patient experience, member experience.”
Key Takeaway:
If your organization is seeking an edge in population health or value-based care, AI-enabled voice analysis could become the next “must-have” capability—shifting mental health upstream and making early intervention routine.
Kintsugi’s technology stands out because it doesn’t require patients or clinicians to change behavior—it layers into conversations already happening. This is rare in a field notorious for workflow disruption and low adoption rates.
“These conversations are happening today... How do we layer in a tool to have those same conversations but maybe surface more actionable information around mental health in real time and track how treatments are going over time?”
Plug-and-play: The AI can be turned on like a “light switch,” embedded in call centers, telehealth, or mobile apps.
Framing for Health Leaders: This isn’t about “throwing away” questionnaires, but about adding another lens—one that’s objective and continuous.
Pappas is candid: “We want you to... do something that's completely different than what currently exists today.”
The future? If voice biomarkers prove out in real-world use, he foresees broad adoption, especially in high-need areas like:
Maternal health (screening for postpartum depression)
Chronic disease management (flagging comorbid mental health risks)
Employee assistance programs (scaling support for large workforces)
“Voice is something that is becoming a little bit more standard of care... The exciting portion is, I get to work with cool clients, but also other cool innovators, to be a very important small part of the overall treatment... helping bridge that for certain MSK or cancer provider.”
Pappas’s career is a case study in building a modern healthcare resume: agility, relationship-building, and deep curiosity about the system’s pain points.
He started out not in technology, but in sports, then education, then pharmaceutical sales—eventually pivoting into medtech, diagnostics, and finally digital health startups.
“One of the hardest things for healthcare... is breaking into it. There's no recipe for that... I was fortunate, and I leveraged my network to figure out what else was out there.”
Network proactively, not transactionally:
“I didn't really have an agenda on the front end... Some of it was, I have a lot to learn in this industry, so would love to learn that way.”
Look for foundational training:
“If you can... have some of those experiences where you get that foundation, then come in [to startups], you can do both.”
Accept non-linear growth:
“From a sales perspective... the framework I've always tried to use is... not to go any steps back. Some of it wasn't a massive financial jump, it was just where I needed to go from a learning curve.”
Don’t mistake titles for substance:
“I kind of joke, call me whatever you want to call me. It usually goes, what stage are we, what are our short-term initiatives, what am I responsible for... chances are it's going to have six or seven things that I'm not on that title to be responsible for.”
Bottom line: The modern healthcare executive must be part operator, part connector, and part relentless learner.
Adopting AI in mental health is not a plug-and-play fix for all systemic challenges—but it offers unique, evidence-driven value where the use case fits.
Kintsugi’s tool shines by augmenting existing processes, not replacing clinicians or dictating workflows. But Pappas cautions that “the flashy and sexy is great, but healthcare isn’t just easy as amazing technology... it still fits in the confines of legacy, dated tech systems.”
Workflow Integration: How easily does the tool layer into your call center, telehealth, or EHR stack?
Patient Experience: Is the tool invisible to the patient, or does it require explicit consent or additional steps?
Staff Training: Will frontline staff need upskilling, or is it truly “frictionless”?
Equity & Privacy: How does the tool perform across diverse populations? What about HIPAA and security?
Proof and Outcomes: Is there published evidence or early outcomes demonstrating cost savings, improved detection, or better engagement?
Early intervention means lower total cost of care.
Objective, continuous data improves risk stratification and population health initiatives.
AI tools can help scale behavioral health support in areas with persistent workforce shortages.
“If your organization is seeking an edge in population health or value-based care, AI-enabled voice analysis could become the next 'must-have' capability—shifting mental health upstream and making early intervention routine.”
The outlook is optimistic—but the real-world impact will depend on thoughtful integration and a focus on outcomes, not just novelty.
Pappas believes adoption will “start with niche use cases—maternal health, chronic disease, call center triage—and expand as evidence mounts.”
“If we can plug it in there, help out a certain area, then we can kind of define what that looks like... the mental health, $200 billion in total cost. The cost is there, the opportunity is there, but it is capturing and making sure that this is something that's frictionless, seamless, adds value.”
Voice is becoming normalized in clinical documentation (AI scribes), patient engagement, and now screening.
AI can bridge the gap for underserved populations—rural, low-access, or stigma-prone communities.
Executives will need to invest not just in technology, but in change management, outcome measurement, and ongoing partnership with solution providers.
For U.S. healthcare executives, the convergence of AI and mental health is not a distant promise—it’s happening now. Kintsugi’s story is a blueprint: focus on embedding technology where it naturally fits, measure impact rigorously, and maintain a bias for learning over perfection. As Pappas says, “If I can even help one or two folks... it just kind of opens up a whole new world of startup and venture capital, how all that happens in digital health.” Leaders who embrace this mindset—and act—will find themselves ahead of the curve, building more resilient, equitable, and patient-centered systems for the years ahead.
For more insights, connect with Josh Pappas on LinkedIn. To learn more about how Kintsugi is reshaping mental health detection, visit their website or follow the American Journal of Healthcare Strategy for future episodes and executive analysis.
