cult.fit, an India-based technology-enabled fitness company operating 708 fitness centers in 77 cities and serving nearly one million paid members, has standardized on OpenAI Codex across engineering and expanded its use into other functions. The company combines subscription access to gyms and trainer-led classes with a growing consumer-products business.
Why the change
As cult.fit scaled, leaders wanted to solve operational and product challenges without adding layers of process. The company sought a repeatable way to move ideas into production, reduce bottlenecks around central teams, and make it easier for different groups to share and reuse software capabilities.
Codex as a common engineering foundation
Before standardizing on Codex, developers relied on a variety of coding agents and assistants, which created inconsistent repository structures, agent instructions, skills and workflows. By choosing Codex, cult.fit established common conventions for repositories, testing, code review and feedback loops across teams.
The company extended Codex into specialized engineering workflows. A Site Reliability Engineering agent aggregates data from observability platforms and logs, investigates incidents and, in many cases, generates a pull request with a proposed fix for an engineer to review. Security agents help detect and remediate vulnerabilities, and other on-call agents investigate recurring production issues.
cult.fit is moving toward a more spec-driven development model in which a specification guides work through the software lifecycle toward a pull request for engineer review. One concrete result: a social feed for the cult.fit app that the CTO Santhosh Kumar estimates would previously have taken more than a year to build was developed end-to-end in two months using the new workflow.
Extending Codex beyond engineering
The company made its internal data easier for Codex to access, enabling teams to find answers and run routine analyses without routing every request through a central analytics team. As a result, non-engineering teams have begun building their own applications and workflows; HR created a Codex workflow to manage performance reviews, scheduling and feedback aggregation. In recent months more than 20 internal applications have been built by non-engineering teams.
When the cost of an internal dashboard software spiked, an engineer rebuilt the functionality with Codex over a weekend. cult.fit now runs that capability in its own application, avoiding more than $100,000 in annual software costs.
Adoption support: hackathons, embedded engineers and office hours
Access alone was not enough. A large hackathon gave employees an early view of what was possible, and a one-day session for management had leaders across functions build relevant solutions with Codex. Engineering holds twice-weekly office hours to help employees move projects from idea to production, and engineers are embedded into selected functions to provide technical and domain support.
AI in member experience and customer support
cult.fit is also using the OpenAI API to enhance member experiences, notably with its AI Trainer. About 45% of cult.fit members are first-time gym-goers, making guidance important to forming a fitness habit. Around 20% of cult.fit’s gym users currently use the AI Trainer product.
The latest AI Trainer version is built around a voice-first experience: using GPT Realtime via an OpenAI API integration, the trainer can provide coaching cues and support workout logging through two-way voice interaction.
In customer support, cult.fit implemented AI agents to handle cases that previously required human assessment each time. For example, when a member requests a penalty waiver for missing a group class, an AI agent can check signals such as local weather and attendance patterns before deciding whether to waive the penalty. These agents are integrated into the company’s Sprinklr support platform and connected to backend systems, shifting support from a mostly human-led chat model toward an AI-first workflow.
Measured impacts and next steps
- More than 20 internal applications have been built by non-engineering teams in recent months.
- Rebuilding an internal capability with Codex avoided more than $100,000 in annual software costs.
- The company reports a greater than 50% reduction in customer support costs.
cult.fit’s next objective is to move from isolated AI workflows to embedding AI into everyday work. CTO Santhosh Kumar describes the ambition as creating an "autopilot operations org," where AI takes on more of the support work around frontline teams. For members, the aim is a "trainer in your pocket," providing scalable, personalized guidance beyond what human trainers alone could provide at scale. Internally, cult.fit continues to search for bottlenecks to remove with AI and to encourage employees to build applications, analyse information and automate workflows themselves.
Summary
By standardizing on Codex and integrating the OpenAI API into products and operations, cult.fit has accelerated delivery, reduced costs, and broadened AI use across engineering, HR, customer support and member experiences. The company plans to continue embedding AI into its operating model so that more ideas can flow directly into working solutions.
