When New York–based healthtech startup Backed by MAYA approached us, they had a bold goal: empower people to prevent and relieve chronic lower back pain with a fully personalized therapy experience—delivered right from their phones. To make it possible, they needed more than a simple fitness tracker or instructional video library. They needed a truly adaptive platform—one that could monitor progress, offer expert guidance, and evolve with each user’s journey. We partnered with them to bring that vision to life, designing and building a smart iOS app supported by an AI engine, progress tracking, and an admin-friendly web platform.
The Problem: Creating Therapy That Evolves with You
Backed by MAYA was built on a clear insight: most people dealing with back pain don’t need surgery or medication—they need consistency, knowledge, and a plan tailored to their condition. The problem? Existing digital solutions were often one-size-fits-all. Most fitness apps offered static routines, lacked clinical oversight, or simply overwhelmed users with information that didn’t apply to their needs.
The team behind MAYA knew there had to be a better way. They envisioned an app that didn’t just show exercises, but actually learned from users’ behavior—adapting in real-time, like a physical therapist would.
To get there, they needed a partner who could:
- Build a sleek and intuitive iOS app
- Develop a backend smart enough to personalize therapy in real time
- Enable therapists or admins to manage content and track progress
- Ensure everything ran securely, reliably, and at scale
Adding to the challenge: this was the startup’s first product launch, and they were working against time to secure investment and release a minimum viable product that demonstrated both clinical value and user engagement.
Our Approach: Marrying Clinical Expertise with Custom Software
From the start, we understood this wasn’t just a “fitness app.” It was a clinical tool disguised as a consumer product—so our team had to balance usability, reliability, and medical integrity.
Discovery and Product Definition
We kicked things off with a focused discovery sprint. Working closely with MAYA’s founders (including clinicians and researchers), we mapped out the user journey from onboarding through daily use. We identified three primary personas: people with recurring back pain, first-time users experiencing strain, and those in post-therapy recovery.
Together, we aligned on the key goals for each user type, including:
- Receiving a customized daily exercise plan
- Tracking progress through feedback (pain levels, effort, results)
- Getting smarter recommendations as therapy continued
Technical Architecture and Stack
To deliver this vision, we chose a stack that prioritized performance and long-term maintainability:
- Mobile App: iOS native app built in Swift, ensuring smooth animations and compatibility with Apple Health integrations.
- Backend: Python-based API and logic layer running an AI engine.
- Web Platform: Admin portal built with React + Node.js, where clinicians could upload content, review user data, and adjust recommendations.
- AI Engine: A lightweight engine trained to match exercises with user profiles and adjust plans based on reported progress and behavior.
The AI engine itself was built to be modular, allowing MAYA’s internal data scientists to iterate on the recommendation logic without needing to recompile or deploy the app every time.
Iterative UX and Visual Design
We designed the app to feel more like a wellness companion than a clinical tool. The daily routine is presented through short, guided video sessions, with calming design elements and intuitive pain tracking. We paid special attention to accessibility—making sure exercises were easy to follow regardless of screen size or user ability.
Users could rate how each session went, and that feedback directly influenced future recommendations. For example, if a user found a session too difficult or painful, the system would adjust the next day’s plan accordingly—reducing intensity or focusing on a different muscle group.
Agile Development and Testing
We worked in two-week sprints, demoing progress to the client weekly. Our QA team conducted ongoing user testing with actual beta users, providing insights into everything from usability to edge cases (e.g., when users skipped sessions or paused mid-workout).
Thanks to this agile approach, we were able to:
- Launch a working MVP in under 4 months
- Iterate rapidly on early feedback
- Build a stable foundation for future expansion (including Android and web)
The Outcome: A Smarter Way to Heal
The result was a digital therapy experience that adapts as users progress. Instead of following rigid workout plans, users receive curated video sessions each day—each one guided by the AI engine and tailored to the user’s recent inputs.
What Users Experience
- A personalized welcome flow that helps define their goals and mobility level
- A daily video routine, typically 5–10 minutes long, with clear voice guidance
- A simple self-assessment tool to rate pain and mobility after each session
- Visual progress tracking that shows how their therapy is evolving
What Clinicians Get
- A centralized admin dashboard to manage exercises, videos, and user data
- Visibility into aggregated user behavior to monitor success and retention
- The ability to tag, update, or replace content without interrupting the app’s operation
What the AI Does
- Matches user profiles with relevant exercise routines
- Adjusts daily plans in real time based on user-reported pain or difficulty
- Detects long gaps or drop-offs in usage and re-engages the user with lighter sessions or reminders
The Results: From Early Adoption to Scalable Potential
Since launch, the app has provided users with a simple, accessible way to engage in daily back care—without needing in-person sessions or specialized equipment. Early adopters have responded positively to the platform’s adaptive design, particularly its ability to evolve exercise routines based on their individual input.
The system’s AI engine allows for continuous personalization, helping users stay engaged and making it easier to maintain consistency over time. The daily video recommendations, intuitive progress tracking, and lightweight onboarding process have contributed to a user experience that feels responsive and supportive.
On the product side, Backed by MAYA now has:
- A fully functional iOS app in the hands of real users
- An AI-powered backend capable of adapting therapy sessions dynamically
- A web-based management platform for monitoring engagement and updating content
- A solid technical foundation for future enhancements, including multi-language support and cross-platform rollout
The early success of the platform has also sparked conversations about broader use cases and potential international expansion—starting with Spanish-speaking markets.
As Ignacio Oroná, our CBO at Zarego, put it:
“Creating something so versatile meant making sure the app could evolve with each user’s unique journey, and that’s what makes it powerful.”
Let’s Talk
Projects like this remind us why we love building software: it’s not just about code—it’s about real-world impact. Whether it’s helping people move without pain or giving startups the tools to scale their vision, we’re here to build smart, sustainable tech that makes a difference.
If you’re working on a health, wellness, or AI-driven product and need a team that can bridge clinical insight with cutting-edge development, let’s talk.
