Recovery Discovery by Show Up and Stay

Recovery Discovery | How a Voice Can Tell a Story (with Henry O’Connell of Canary Speech)

DeAnn Knighton Season 4 Episode 12

What if a short snippet of conversation could reveal insights into your emotional, behavioral, and cognitive health and signal risk factors for Substance Use Disorder? In this episode of Recovery Discovery, hosts' DeAnn Knighton and Craig Knighton sit down with Henry O’Connell, Co-Founder and CEO of Canary Speech, to explore how artificial intelligence and voice biomarkers are transforming the way clinicians understand human well-being.

From its origin story in a Utah bagel shop to its deployment in primary care and behavioral health settings, Canary Speech’s “ambient listening” technology captures subtle, subconscious vocal patterns — helping clinicians identify indicators of anxiety, depression, pain, mild cognitive impairment, and even Alzheimer’s disease.
The three discuss:

  • The science of voice biomarkers and clinical decision support
  • How Canary’s non-intrusive AI assists rather than replaces clinicians
  • Application to substance use disorder 
  • Real stories from patients and practitioners using the tool
  • The ethical and emotional dimensions of integrating AI into mental health care and substance use disorder

This conversation is equal parts innovation, empathy, and inspiration — a look at how technology can help us listen better to ourselves and one another

If you have questions or comments or would like to be a guest on the Recovery Discovery Podcast, send an email to:
info@showupandstay.org

More information about our project is available at:
https://www.showupandstay.org/
https://www.soberpositiveworkplace.org/

For podcast updates, follow us on Instagram:
https://www.instagram.com/showupandstayorg/


Music and Audio Production by Katie Hare.
https://www.hare.works

When Voices Tell the Story: How Canary Speech Is Transforming Behavioral Health and Addiction Recovery Through AI

By DeAnn Knighton | Recovery Discovery Podcast

What if the sound of your voice could tell a clinician how you’re really doing—without a single question on a form?

That’s the world being built by Henry O’Connell, CEO and Co-Founder of Canary Speech, a pioneering company using AI voice biomarkers to transform how clinicians assess behavioral, cognitive, and addiction-related health. In this episode of Recovery Discovery, Henry and I explore how a conversation in a Utah bagel shop evolved into a breakthrough technology now helping healthcare professionals listen more deeply—especially in the context of substance use and recovery.

From Bagels to Breakthroughs

The story begins with friendship and curiosity. Over lunch with longtime colleague Jeff Adams, the speech and language innovator behind Dragon NaturallySpeaking, Nuance, and the core technology that became Amazon Alexa, Henry posed a question:

“What if we could use speech and language to understand human health and disease?”

That conversation planted the seed for Canary Speech—a platform that uses AI to analyze the acoustic and linguistic features of human voice to identify biomarkers of emotional, behavioral, and neurological change.

The concept was rooted in empathy. Henry recalled noticing, even over the phone, when his daughter Caitlin was struggling or stressed: “It had nothing to do with the words she was saying. I could just hear it.”

If parents and clinicians intuit emotion through tone and rhythm, could technology learn to do the same—objectively and at scale?

Voice as a Mirror for the Recovery Journey

In addiction treatment, much of the clinical process depends on self-reporting—tools like the PHQ-9 or GAD-7, which rely on an individual’s perception of their own mood or anxiety levels. These measures are valuable but subjective, and many patients—especially those in early recovery—struggle to articulate what they feel.

That’s where Canary Speech enters the conversation.

By analyzing subtle changes in pitch, pace, micro tremors, pauses, and breath patterns, the AI can detect signs of stress, craving, agitation, or cognitive fatigue—often before the person themselves is consciously aware of it.

“Canary Speech isn’t replacing the clinician,” Henry explained. “It’s reinforcing what clinicians already know to look for—only now, they can validate it with objective data.”

This opens powerful new pathways in addiction medicine, telehealth, and aftercare, where continuous listening could flag early relapse risk, withdrawal-related distress, or co-occurring anxiety and depression—all without additional burden on the clinician or patient.

How It Works: AI That Listens Differently

Canary’s ambient listening model means the patient doesn’t have to “take a test.” Instead, the system analyzes voice data during natural conversation—whether that’s a telehealth session, a routine follow-up, or a peer-support interaction.

The software processes over 2,500 biomarkers, mapping how the central nervous system translates emotion into sound. Because these biomarkers are subconscious—outside voluntary control—they provide a uniquely reliable measure of internal state.

“It doesn’t matter whose hands the tool is in,” Henry said. “It’s going to respond the same way—because the data is biological, not behavioral.”

For addiction recovery teams, that means real-time feedback on whether a patient’s emotional regulation, motivation, or stress load is shifting—critical indicators of relapse vulnerability.

A Companion, Not a Replacement

In our discussion, Henry described Canary as a “companion in the toolbox”—a supportive presence that complements, not replaces, the therapeutic alliance.

He likens it to a Winnie-the-Pooh analogy: “Some of us are Tiggers, some are Eeyores. Canary isn’t judging either—it’s helping clinicians recognize who’s in front of them that day.”

For addiction professionals, that insight is invaluable. Recovery isn’t linear; emotional tone can fluctuate long before behavior does. Canary gives clinicians a quiet assist in noticing those changes early—and responding with care rather than crisis management.

Real Stories, Real Impact

Henry shared one striking example: a primary care team using Canary detected early signs of mild cognitive impairment in a 95-year-old patient during a routine visit. That early detection connected the family with Alzheimer’s support, counseling, and resources.

In addiction treatment, similar early detection could translate into preventing relapse, flagging burnout in recovery programs, or supporting post-treatment monitoring—where consistent engagement is key, and silent distress often goes unnoticed.

The Future of Listening in Recovery and Mental Health

As artificial intelligence reshapes clinical care, tools like Canary Speech stand out for one reason—they make listening the technology’s central function.

Rather than diagnosing, it illuminates patterns. Rather than replacing intuition, it validates connection. And in the context of addiction recovery, that combination—human empathy enhanced by data—may be one of the most promising advances yet.

“Technology shouldn’t replace intuition,” Henry told us. “It should validate it.”

That’s the power of Canary Speech—and the future of recovery care that listens not just to what we say, but how we say it.

🎧 Listen to the full episode: How a Voice Can Tell a Story: AI, Emotion, and the Future of Behavioral Health with Henry O’Connell
Available now on the Recovery Discovery Podcast.

Keywords: AI in Healthcare, Behavioral Health Tech, Addiction Recovery, Voice Biomarkers, Clinical Innovation, Recovery Discovery Podcast, Canary Speech, Mental Health AI, Neuro Health