The Center for Data Innovation recently spoke with Mike Telem, co-founder of Kemtai, an Israel-based company developing an app that enables patients to complete prescribed exercises at home while receiving real-time corrective guidance. Telem explained how the platform uses computer vision to track patients’ movements as they complete exercises and translates those movements into actionable insights for providers.
David Kertai: Why do providers struggle to deliver effective home-based physical therapy?
Mike Telem: In home-based physical therapy, clinicians typically prescribe exercises during visits at a clinic or doctor’s office, then ask patients to complete those exercises independently at home. While instructions are often provided verbally, in writing, or sometimes demonstrated during the visit, there’s limited feedback once patients leave and little visibility into whether exercises are performed correctly or consistently. Patients may follow instructions as best they can, but mistakes in form or skipped sessions can slow recovery, and clinicians have no objective way to track progress.
Our app helps close this gap by turning everyday devices like phones or laptops into guided therapy tools. At home, patients follow prescribed exercises while the system tracks movements in real-time, providing immediate feedback on posture, alignment, and technique. All movement data is automatically uploaded to the cloud, where clinicians can view structured reports, reconstructed motion replays, and performance metrics. This enables providers to adjust care plans based on actual performance rather than self-reported adherence, making home exercises safer, more consistent, and measurable.
Kertai: How does your computer vision technology track patient movement?
Telem: We developed a motion-tracking software system that runs through a mobile app or online website and works directly through a standard camera without requiring wearables or specialized hardware. Once the patient opens the app, it guides them through prescribed exercises while capturing movements. The software identifies key body points—such as elbows, wrists, and knees—analyzes joint angles, posture, and overall movement quality, and uploads structured data to the cloud.
To maintain accuracy in home environments, the system works across different devices, body types, clothing, and lighting conditions. It continuously calibrates during each session and uses motion tracking to reliably follow key areas, such as hips and knees. This ensures patients receive real-time guidance and clinicians can see how exercises are performed, helping make home-based physical therapy more effective.
Kertai: How does your platform translate that movement data into accurate, real-time feedback?
Telem: As patients exercise, the platform evaluates each repetition and compares it to the intended movement pattern. It extracts metrics such as range of motion, stability, and symmetry, then uses that data to deliver immediate corrective cues, such as adjusting posture or modifying arm position.
At the same time, the system aggregates this information into detailed reports for clinicians. These reports include session-by-session performance data and reconstructed movement replays that show how a patient moved. This approach gives patients clearer guidance and accountability while helping clinicians better understand progress between visits.
Kertai: How does Kemtai personalize care while maintaining scalability?
Telem: Clinicians create care plans using configurable templates that they tailor to each patient’s condition, mobility level, and goals. These plans evolve over time, either on a set schedule or based on what the patient completes. A plan can include specific exercises, periodic checks on a patient’s range of motion, and simple patient-reported check-ins on pain or difficulty. This structure lets clinicians deliver individualized care to many patients at once, making the experience personal without sacrificing efficiency.
Kertai: What challenges have you faced bringing this technology into the physical rehab space?
Telem: One of the biggest challenges was earning clinicians’ trust. Physical therapists rely heavily on hands-on assessment and professional judgment, so we needed to demonstrate that technology could effectively support their work.
We addressed this by prioritizing scientific validation to demonstrate accuracy and reliability, co-designing the platform with therapists to ensure it fit real clinical workflows, and keeping the system simple for both patients and providers. By involving clinicians throughout development, we built a tool that integrates naturally into practice rather than disrupting it.
