Grounded in science. Powered by empathy.

Ebby™ combines behavioral science modeling, biometric data, and predictive AI to deliver real-time, personalized support during recovery right when it’s needed most.

The Problem

Recovery needs a smarter system

Despite millions entering treatment each year, relapse rates remain stubbornly high, exceeding 60% in the first year alone. Recovery today is reactive—support often vanishes right when it’s needed most.

Most digital tools rely on generic check-ins and delayed responses. But relapse isn't generic; it's personal and unpredictable, quietly building across behaviors, moods, and environments until it becomes overwhelming.

Ebby™ intervenes before relapse takes hold, providing proactive support exactly when it counts. Because recovery deserves smarter, faster, human-centered support.

Scientific Foundations

Ebby™ draws on interdisciplinary insights from:

Behavioral science and relapse prevention theory

Behavioral science and relapse prevention theory

Behavioral science and relapse prevention theory

Behavioral science and relapse prevention theory

Digital phenotyping and passive monitoring

Digital phenotyping and passive monitoring

Digital phenotyping and passive monitoring

Digital phenotyping and passive monitoring

AI-guided intervention systems

AI-guided intervention systems

AI-guided intervention systems

AI-guided intervention systems

Clinical research into sleep, pain, and substance use dynamics

Clinical research into sleep, pain, and substance use dynamics

Clinical research into sleep, pain, and substance use dynamics

Clinical research into sleep, pain, and substance use dynamics

Human-centered design in digital mental health

Human-centered design in digital mental health

Human-centered design in digital mental health

Human-centered design in digital mental health

Our models are informed by the latest research and continuously refined through structured pilot studies and user feedback.

Predictive Relapse Prevention (PRP)

At the heart of RecoveryAI’s platform is our Predictive Relapse Prevention (PRP) engine: a proprietary, adaptive system that analyzes real-world behavioral signals to predict relapse risk before it escalates.

Real-Time Behavioral Monitoring

Ebby™ ingests mood signals, journaling tone, usage patterns, biometric data, and geolocation to build a continuous, real-time risk profile.

Real-Time Behavioral Monitoring

Ebby™ ingests mood signals, journaling tone, usage patterns, biometric data, and geolocation to build a continuous, real-time risk profile.

Real-Time Behavioral Monitoring

Ebby™ ingests mood signals, journaling tone, usage patterns, biometric data, and geolocation to build a continuous, real-time risk profile.

Real-Time Behavioral Monitoring

Ebby™ ingests mood signals, journaling tone, usage patterns, biometric data, and geolocation to build a continuous, real-time risk profile.

Dynamic Risk Scoring

Our model calculates personalized relapse risk scores that update dynamically as new data is received.

Dynamic Risk Scoring

Our model calculates personalized relapse risk scores that update dynamically as new data is received.

Dynamic Risk Scoring

Our model calculates personalized relapse risk scores that update dynamically as new data is received.

Dynamic Risk Scoring

Our model calculates personalized relapse risk scores that update dynamically as new data is received.

Just-in-Time Adaptive Interventions (JITAI)

Based on real-time risk scores, Ebby™ delivers personalized, evidence-based interventions exactly when they are most needed.

Just-in-Time Adaptive Interventions (JITAI)

Based on real-time risk scores, Ebby™ delivers personalized, evidence-based interventions exactly when they are most needed.

Just-in-Time Adaptive Interventions (JITAI)

Based on real-time risk scores, Ebby™ delivers personalized, evidence-based interventions exactly when they are most needed.

Just-in-Time Adaptive Interventions (JITAI)

Based on real-time risk scores, Ebby™ delivers personalized, evidence-based interventions exactly when they are most needed.

Continuous Learning

The PRP system refines and adapts its predictive models over time, learning from every user interaction to improve precision and personalize support.

Continuous Learning

The PRP system refines and adapts its predictive models over time, learning from every user interaction to improve precision and personalize support.

Continuous Learning

The PRP system refines and adapts its predictive models over time, learning from every user interaction to improve precision and personalize support.

Continuous Learning

The PRP system refines and adapts its predictive models over time, learning from every user interaction to improve precision and personalize support.

The Difference

Why RecoveryAI’s intelligent recovery system is different

Built for prediction, not reaction

RecoveryAI identifies early warning signs using behavioral, emotional, and contextual signals, intervening proactively before relapse risk escalates.

Built for prediction, not reaction

RecoveryAI identifies early warning signs using behavioral, emotional, and contextual signals, intervening proactively before relapse risk escalates.

Built for prediction, not reaction

RecoveryAI identifies early warning signs using behavioral, emotional, and contextual signals, intervening proactively before relapse risk escalates.

Built for prediction, not reaction

RecoveryAI identifies early warning signs using behavioral, emotional, and contextual signals, intervening proactively before relapse risk escalates.

Personalized in the moment

Support is tailored dynamically to each individual's behaviors, needs, and environment.

Personalized in the moment

Support is tailored dynamically to each individual's behaviors, needs, and environment.

Personalized in the moment

Support is tailored dynamically to each individual's behaviors, needs, and environment.

Personalized in the moment

Support is tailored dynamically to each individual's behaviors, needs, and environment.

Continuously adaptive

Our system learns from every interaction, improving precision, reducing false positives, and evolving alongside the user's recovery journey.

Continuously adaptive

Our system learns from every interaction, improving precision, reducing false positives, and evolving alongside the user's recovery journey.

Continuously adaptive

Our system learns from every interaction, improving precision, reducing false positives, and evolving alongside the user's recovery journey.

Continuously adaptive

Our system learns from every interaction, improving precision, reducing false positives, and evolving alongside the user's recovery journey.

Grounded in evidence

Intervention pathways are built on validated frameworks including CBT, DBT, Motivational Interviewing, and relapse prevention theory, delivered intelligently and at scale.

Grounded in evidence

Intervention pathways are built on validated frameworks including CBT, DBT, Motivational Interviewing, and relapse prevention theory, delivered intelligently and at scale.

Grounded in evidence

Intervention pathways are built on validated frameworks including CBT, DBT, Motivational Interviewing, and relapse prevention theory, delivered intelligently and at scale.

Grounded in evidence

Intervention pathways are built on validated frameworks including CBT, DBT, Motivational Interviewing, and relapse prevention theory, delivered intelligently and at scale.

We're addressing the whole recovery ecosystem.

RecoveryAI looks beyond cravings and mental triggers. True recovery is shaped by real-world barriers like housing instability, transportation gaps, and food insecurity—factors that quietly escalate risk over time.

Our platform integrates Health-Related Social Needs (HRSN) pathways to connect users with essential resources when they need them most.

Meet The Science Team

Joe Kossowsky

Co-Founder & Head of Scientific Strategy

Joe Kossowsky is an Assistant Professor at Harvard Medical School and a Clinical Researcher at Boston Children’s Hospital. He is a clinical psychologist and biomedical informatician who specializes in integrating wearable device data, developing adaptive intervention frameworks, and creating evidence-based strategies for substance use and behavioral health. At RecoveryAI, Kossowsky leads scientific validation efforts, clinical strategy, and ensures the ethical implementation of our technologies.

Joe Kossowsky

Co-Founder & Head of Scientific Strategy

Joe Kossowsky is an Assistant Professor at Harvard Medical School and a Clinical Researcher at Boston Children’s Hospital. He is a clinical psychologist and biomedical informatician who specializes in integrating wearable device data, developing adaptive intervention frameworks, and creating evidence-based strategies for substance use and behavioral health. At RecoveryAI, Kossowsky leads scientific validation efforts, clinical strategy, and ensures the ethical implementation of our technologies.

Joe Kossowsky

Co-Founder & Head of Scientific Strategy

Joe Kossowsky is an Assistant Professor at Harvard Medical School and a Clinical Researcher at Boston Children’s Hospital. He is a clinical psychologist and biomedical informatician who specializes in integrating wearable device data, developing adaptive intervention frameworks, and creating evidence-based strategies for substance use and behavioral health. At RecoveryAI, Kossowsky leads scientific validation efforts, clinical strategy, and ensures the ethical implementation of our technologies.

Joe Kossowsky

Co-Founder & Head of Scientific Strategy

Joe Kossowsky is an Assistant Professor at Harvard Medical School and a Clinical Researcher at Boston Children’s Hospital. He is a clinical psychologist and biomedical informatician who specializes in integrating wearable device data, developing adaptive intervention frameworks, and creating evidence-based strategies for substance use and behavioral health. At RecoveryAI, Kossowsky leads scientific validation efforts, clinical strategy, and ensures the ethical implementation of our technologies.

From consumer product to digital therapeutic

RecoveryAI is on a dual-track path:
Start simple and prove impact.

Immediate access through a direct-to-consumer model

Build toward clinical-grade care.

Long-term validation and FDA clearance as a reimbursable DTx

We are actively applying for NIH/NIDA funding and working with leading addiction treatment providers to validate impact across:

Relapse rates

Relapse rates

Relapse rates

Relapse rates

Recovery Capital

Recovery Capital

Recovery Capital

Recovery Capital

Health-related social outcomes

Health-related social outcomes

Health-related social outcomes

Health-related social outcomes

Long-term engagement and system scalability

Long-term engagement and system scalability

Long-term engagement and system scalability

Long-term engagement and system scalability

Want to be part of what we’re building?

Join the early access waitlist and experience the most advanced relapse prevention system ever built.

Want to be part of what we’re building?

Join the early access waitlist and experience the most advanced relapse prevention system ever built.

Want to be part of what we’re building?

Join the early access waitlist and experience the most advanced relapse prevention system ever built.

Want to be part of what we’re building?

Join the early access waitlist and experience the most advanced relapse prevention system ever built.

© 2025 Recovery AI, Inc. All rights reserved.

© 2025 Recovery AI, Inc. All rights reserved.

© 2025 Recovery AI, Inc. All rights reserved.

© 2025 Recovery AI, Inc. All rights reserved.