october 2024 - july 2025
Netherlands
sport, fitness, AI
- 1 RoR developer
- 1 PM
challenge
unstable legacy system
The client inherited a system filled with bugs and architectural flaws. Poor code quality and lack of documentation made it difficult to maintain, extend, or even fully understand how the platform worked.
missing core functionality
Essential features for member engagement, such as segmentation, activity monitoring, and lead management, were absent. This limited the platform’s ability to deliver value to gym owners and members.
no user adoption
With no active clients or pilot gyms onboard, the product had not yet proven its viability in the market. Before attracting customers, the system had to be stabilized, modernized, and equipped with competitive features.
demand for AI-driven capabilities
To differentiate in a competitive fitness tech landscape, the client needed advanced AI solutions, including a virtual trainer, a receptionist for handling inquiries, and smart lead-generation logic to attract and retain new members.
solution
stabilize the legacy system
We began by fixing critical bugs across the platform, including message delivery errors in WhatsApp integration and multiple frontend/backend issues. This step was crucial to restore system stability and prepare the foundation for future feature development.
build advanced member management features
To help gyms better engage with their clients, we developed member segmentation based on visiting behavior, added bulk uploading via XLSX files, and significantly improved the journeys system. The latter was enhanced with new step types, stronger validations, and higher fault tolerance.
integrate AI-powered assistantst
Our team implemented two key AI features:
- 1. AI receptionist – an OpenAI-based WhatsApp bot capable of answering questions about gym hours, classes, and equipment.
- 2. AI trainer – a virtual coach available 24/7 to provide personalized fitness advice.
We also enabled document attachment for extending the bots’ knowledge base and introduced an admin/AI switcher to let staff seamlessly take over conversations when needed.
enable smart lead management
To support customer acquisition, we created a dedicated interface and AI logic for handling conversations with potential members. The AI was trained to highlight gym advantages, sustain interest, and encourage leads to join. It also detects open questions or strong buying signals and automatically alerts administrators via email for timely follow-up.
modernize the frontend
The user interface was reworked to align with the new design, which ensures a more polished and intuitive user experience.
strengthen system reliability
We increased test coverage by introducing unit tests for all newly developed features. This improved development confidence, reduced regressions, and ensured higher product stability.
technology stack
Backend
Ruby Ruby on Rails Postgres
Frontend
Turbo Stimulus
Integrations
Dialog360 OpenAI API Postmark Bugsnag
Development & Deployment
Git RubyMine Render
communication
We managed project tasks and tracked progress using Asana. Slack served as the main channel for day-to-day communication with the client. Weekly Google Meet calls were initially held to discuss priorities, clarify requirements, and resolve complex issues. As the client preferred minimizing calls, these meetings were gradually replaced by detailed asynchronous updates in Slack. This approach allowed the team to maintain efficiency across different time zones and keep the client informed without unnecessary overhead.
Estimate your project
Please fill out this form, and our manager will contact you within one business hour. If necessary, we can sign an NDA and begin project discussions.
Thank you.
Your message has been sent successfully!
We’ll get in touch with you within 24 hours, excepting requests received on Saturday, Sunday.