Injury Rehabilitation Reimagined

Problem Space/Brief:The challenges facing athletes recovering from injuries are multifaceted. Existing rehabilitation methods lack personalization, accessibility, and efficiency, often resulting in prolonged recovery times, increased risk of re-injury, and dissatisfaction among practitioners. The absence of real-time monitoring and adjustment of recovery plans, coupled with limited access to specialized services, further compounds these challenges.

Without personalized support and feedback, people struggle to maintain motivation and adherence to rehabilitation protocols, leading to decreased performance and increased healthcare costs. Developing an AI-powered rehabilitation coacch presents an opportunity to address these challenges by providing personalized, adaptive support that optimizes recovery outcomes and enhances user satisfaction and engagement throughout the rehabilitation process.

View Methodology
View Solution

Contextual Inquiry

METHOD

Shadowed participants during their sessions and conducted informal interviews to gather insights in context.

Participants

3 Physical Therapists
1 Physical Therapy Assistant
1 Student Intern

Key Findings

01

Personalized products are manual and time-intensive

Therapistys tailor exercises based on injury type, patient feedback, and performance, Mostly relying on memory, paper records, or EMRs

02

REAL time form correction is critical

Therapists constantly correct form to prevent injury and maxamise effectiveness. Assistants and interns noted that patients often struggle to replicate exercises corredctly at home.

03

Progress monitoring lacks granularity

Documentation is often updated sporadically, limiting the ability to track subtle improvements or setbacks between rehabilitation sessions.

04

Patient engagement is inconsistent

Motivation varies widely. Patients often forget instructions or lose interest in repetitive routines.

05

Workflow interuiptions are frequent

Staff Juggle administrative tasks, hands on therapy, asnd coaching, leaving little to no time for deep data entry or follow ups.

Implications for Design

AI Recomendation Engine: Must mirror how theapists think: Adaptable, Injury specific, and eresponsive to realtime guidance therapists give.

Form Feedback: Vision-based feedback tools can replicate the real-time guidance therapists give

Progress Insights: Daily session based tracking that quantifies micro improvements could bridge gaps between in clinic and at home therapy.

Patient UX: Gamification or motivational prompts based on performance data could improve adherence.

Application workflow

A step‑by‑step walkthrough of how the AI‑powered rehabilitation coach integrates into home recovery routines.
‍Onboarding & Injury Profiling – Upon first use, athletes complete a brief form detailing injury type, history, and current limitations. The AI uses this to generate a customized rehab plan.
‍In‑Clinic Training Session – With therapist supervision, the system introduces exercises. AI vision tracks form in real time, giving immediate feedback such as “bend deeper” or “shorten range of motion.”
‍Home‑Based Sessions – Athletes continue their routines at home guided by the app, with on‑screen form corrections and motivational prompts.
‍Progress Sync & Review – Data from each session syncs to a clinician dashboard. Therapists review metrics like joint angle consistency and rep counts before next session.
‍Plan Adjustment & Feedback Loop – Therapists can tweak routines in the system, and AI dynamically adjusts difficulty levels. Notifications inform athletes of updates.

Conceptual Wireframes

Wireframes serve as a foundational step in the design process, transforming early ideas into structured, low-fidelity layouts. They help visualize information architecture, user flow, and core interactions without getting distracted by visuals or branding. In this project, wireframes were used to align the team on functionality and layout decisions before moving into high-fidelity UI and prototyping stages.

Data COnsiderations

Data plays a critical role throughout the product lifecycle.. In this project, we identified key data points early on to ensure the AI system could deliver personalized rehab insights, while also supporting clinicians with actionable metrics. These considerations influenced both the technical architecture and the UX, guiding decisions around privacy and interface design.

Solution

The solution combines AI-driven assessment tools with a personalized rehabilitation assistant to support athletes throughout their recovery journey.
‍
AI-Assisted Assessment Tools
Using movement tracking and input from wearable or camera data, these tools evaluate range of motion, strength, and form. They give therapists and athletes actionable insights — helping detect imbalances, measure progress, and fine-tune regimens in real time.
‍
Personalized Recovery Assistant
An interactive AI guide recommends PT-approved exercises based on injury type, recovery phase, and user feedback. It delivers real-time coaching on form, keeps motivation high with progress tracking, and adapts to each athlete’s evolving needs.Together, these tools empower both therapists and athletes to make informed decisions — turning recovery into a data-supported, user-centered experience.

I designed an AI assistant that delivers personalized exercise regimens, real-time plan updates, and progress tracking; all tailored to the unique needs of recovering athletes. Through uncovering insights from physical therapists, the solution simplifies the rehab journey through these key experiences:

AI aSSISTED Assessment Tools

Automatically analyze form, range of motion, and posture using AI-powered pose detection using convolutional neural networks . These tools give athletes instant feedback and provide therapists with better insights to serve clients.

Progress Tracking

A simple, clean dashboard gives athletes a clear view of their recovery journey that includes completed assessments, updated plans, and tracked metrics like pain level, range of motion, and strength. This visibility helps users stay motivated and informed as they work toward recovery goals.

Smart Exercise Matching

Powered by recommendation algorithms trained on PT routines and injury profiles, this system uses reinforment learning to defined the best exercises based on recovery stage, movement history, and therapist input. As the user improves, the assistant adapts their plan using reinforcement from prior outcomes.

Real-Time Feedback

Using a combination of sensor data (e.g. motion, heart rate) and user-reported feedback, the assistant dynamically adjusts the recovery plan in real time. A lightweight machine learning model continuously evaluates whether the current exercise intensity or sequence is effective then recommends changes if anomalies are detected (e.g. overexertion, underperformance).