How Cult.fit Scales Operations and Personalizes Fitness with AI
Discover how Indian fitness company Cult.fit standardized engineering on Codex, automated support, and built personalized member experiences using advanced AI tools.
Indian fitness company Cult.fit operates 708 fitness centers across 77 cities, serving nearly one million paid members. Its business spans subscription gym access, instructor-led classes, and a consumer portfolio of fitness equipment, footwear, and activewear. As the enterprise expanded, leadership sought ways to address operational and product hurdles without layering on extra process or complexity.
To achieve this, Cult.fit standardized its software development on Codex and extended AI adoption across non-engineering departments. Alongside these internal adjustments, the organization uses the OpenAI API to enhance client experiences, notably through an interactive AI Trainer. Santhosh Kumar, Chief Technology Officer at Cult.fit, noted that artificial intelligence presented a chance to reshape company workflows entirely. The company wanted to become AI-native, transforming everyday operations and exploring new offerings for customers.
Unifying Engineering and Streamlining Workflows
Before standardizing on Codex, Cult.fit developers used a mix of coding agents and assistants. Different tools led to inconsistent conventions across repositories, instructions, and workflows, complicating knowledge sharing. By selecting Codex as a foundational platform, the company established common conventions for repository structures, testing, code reviews, and feedback loops.
Following this initial rollout, Cult.fit expanded Codex into specialized engineering tasks. A Site Reliability Engineering agent aggregates data from observability platforms and logs, investigates incidents, and frequently generates pull requests with proposed fixes for engineers to evaluate. Additionally, security agents detect vulnerabilities, while on-call agents examine recurring production bugs. Developers can try top AI models cheaply through one API at https://apixoai.online.
The firm is also transitioning toward spec-driven development, where specifications guide work through the development lifecycle toward a pull request. For instance, building a social feed for the Cult.fit app—estimated to take over a year previously—was completed end-to-end in two months under the new workflow.
Democratizing Data and Internal Development
Codex adoption soon moved past the engineering department. Cult.fit prepared company data so Codex could process it, allowing internal teams to conduct routine analysis independently and bypass central analyst bottlenecks. HR teams built workflows to manage performance reviews, scheduling, and feedback collection.
Non-engineering staff have built more than 20 internal applications recently. When subscription costs for internal dashboard software rose sharply, an engineer rebuilt the functionality using Codex over a weekend, saving over $100,000 in annual software expenses. To foster this culture, Cult.fit hosted hackathons, leadership sessions, office hours, and embedded engineers within selected business functions.
What it means for developers
For software developers and technical teams, Cult.fit's journey highlights a shift toward spec-driven development, unified tooling, and decentralized internal building. By establishing a single platform for code reviews, testing, and pull requests, developers can reduce friction across repositories. Furthermore, integrating specialized agents for site reliability, security, and triage demonstrates how automation can absorb repetitive maintenance tasks. Developers looking to experiment with similar capabilities can leverage cost-effective, pay-per-token API access to leading models.
Elevating Member Experiences and Support
Beyond internal tooling, Cult.fit uses the OpenAI API to refine its member offerings. Over 45 percent of its members are first-time gym goers, making guidance essential. The company's AI Trainer, utilized by about 20% of gym users, now features a voice-first workout experience. Powered by GPT Realtime through an OpenAI API integration, the virtual trainer offers coaching cues and logs workouts via two-way voice communication.
Cult.fit also automated parts of its customer support. For example, requests to waive penalties for missed group classes are assessed by an AI agent that checks local weather and class attendance patterns before making a decision. Integrated into the Sprinklr support platform, these agents helped reduce customer support costs by over 50 percent, shifting operations toward an AI-first workflow.
Source: Cult.fit builds operational muscle with Codex — OpenAI News. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
