Cult.fit Enhances Operations and Member Experience with Codex and AI

Cult.fit, a leading technology-enabled fitness company in India, operates 708 fitness centers across 77 cities and serves nearly one million paid members. The company combines subscription-based access to gyms and trainer-led fitness classes with a growing portfolio of consumer products, including fitness equipment and activewear.

As Cult.fit expanded, it sought to address operational challenges without adding complexity. By standardizing its engineering processes on Codex, the company enables teams to analyze data, build tools, and implement ideas efficiently. Additionally, Cult.fit utilizes the OpenAI API to create personalized member experiences, including its AI Trainer.

Santhosh Kumar, CTO of Cult.fit, views AI as a transformative opportunity for operations. He stated, “We wanted to become more AI-native in the way we run things. How can we run our operations or transform our operations to be AI-first?”

Standardizing Software Development with Codex

The major shift for Cult.fit began with its engineering practices. Before adopting Codex, developers used various coding agents, leading to inconsistent conventions and challenges in knowledge sharing. The decision to standardize on Codex established a unified platform for the software development lifecycle, ensuring consistent repository structures and workflows.

Kumar noted, “Once we standardized on Codex, we could build our software development lifecycle around a single platform.” The team extended Codex into specialized workflows, such as Site Reliability Engineering, which aggregates information from observability platforms and logs, and security agents that identify vulnerabilities.

Cult.fit is also transitioning to a specification-driven development model, guiding work toward pull requests for review. A recent example includes the development of a social feed for the Cult.fit app, which was completed in two months using the new workflow, a process that previously took over a year.

Empowering Teams to Build

Codex’s capabilities have expanded beyond engineering, making company data more accessible for team analyses. Previously, minor data requests created bottlenecks as they had to go through a central analyst team. Kumar remarked, “Codex has been our biggest step toward data democratization. Teams can now do the analysis they need themselves.”

Various teams, including HR, have developed workflows in Codex, resulting in over 20 internal applications created by non-engineering teams in recent months. This shift has altered Cult.fit’s approach to necessary software, enabling engineers to rebuild internal dashboard functionality over a weekend, saving the company over $100,000 annually.

Implementing this change involved more than just access to Codex; a large hackathon allowed employees to explore its potential, followed by sessions for management to build relevant applications. Engineering also hosts bi-weekly office hours to assist employees in moving projects from concept to production.

Enhancing Member Experience with AI

Cult.fit utilizes the OpenAI API to enhance member experiences through its AI Trainer. With over 45% of members being first-time gym-goers, guidance is essential for establishing fitness routines. The AI Trainer, developed several years ago, provides personalized workout experiences without a personal trainer.

Currently, around 20% of gym users utilize the AI Trainer, which has evolved into a more interactive experience using GPT Realtime through OpenAI API integration for coaching cues and workout logging via voice interaction.

Integrating AI into Operations

Cult.fit applies AI to customer support workflows that previously required human agents. For example, when members request to waive penalties for missed classes, an AI agent evaluates the request by analyzing local weather and attendance patterns before making a decision.

This integration has shifted customer support from a human-led chat model to an AI-first workflow, enhancing efficiency.

Results at a Glance

• 20+ internal applications built by non-engineering teams
• $100,000+ annual software savings by rebuilding an internal capability with Codex
• 50%+ reduction in customer support costs

Future Directions

Cult.fit aims to evolve from individual AI workflows to an organization where AI is integrated into everyday operations. Kumar envisions an “autopilot operations org,” where AI supports frontline teams. The goal is to create a “trainer in your pocket,” providing personalized guidance at scale. As more employees create applications and automate workflows, greater ideas can transition into functional solutions.

Source: OpenAI News