Cult.fit operational transformation with OpenAI Codex

Standardizing software development with Codex

Cult.fit has standardized its engineering lifecycle on Codex to eliminate fragmentation caused by using multiple disparate coding assistants. By centralizing repository structures, testing, and code review workflows under a single platform, the company has achieved more consistent development cycles. This standardization extends to specialized engineering tasks, including Site Reliability Engineering (SRE) agents that monitor logs, investigate incidents, and generate pull requests for fixes. The company is also shifting toward a spec-driven development model, which has significantly accelerated delivery; for instance, a social feed feature that previously would have required over a year to build was completed and shipped in two months.

Democratizing internal application development

Beyond engineering, cult.fit has used Codex to empower non-technical teams to build their own tools and analyze data, effectively removing bottlenecks that previously required central analyst support. HR teams have utilized the platform to automate performance review processes, and employees have collectively built over 20 internal applications in recent months. This shift has also yielded direct financial benefits; one engineer replaced a costly third-party internal dashboard with a custom solution built via Codex over a single weekend, resulting in over $100,000 in annual software savings. To support this adoption, the company hosts twice-weekly office hours and embeds engineers within non-technical functions to assist with project development.

Enhancing member experience with AI

Cult.fit is integrating the OpenAI API to evolve its "AI Trainer" product, which provides personalized workout guidance to members. Approximately 20% of the company's gym users currently utilize this tool. The latest iteration incorporates GPT Realtime to enable a voice-first experience, allowing for two-way conversational coaching and automated workout logging. This initiative aims to provide scalable, personalized fitness guidance to the company's large member base, 45% of whom are first-time gym attendees.

Automating operational workflows

AI is being integrated into cult.fit’s customer support infrastructure to move from human-led chat models to AI-first workflows. By connecting AI agents to the company's existing Sprinklr support platform and backend systems, the company can now automate complex decision-making tasks, such as evaluating requests to waive penalties for missed fitness classes. These agents assess requests by analyzing contextual data like local weather and class attendance patterns, contributing to a reported 50% reduction in customer support costs.

Future outlook

Cult.fit aims to transition toward an "autopilot operations" model where AI handles the majority of support tasks for frontline teams. The company continues to prioritize the integration of AI into its core operating model, with the goal of scaling personalized member guidance through a "trainer in your pocket" concept while further reducing organizational bottlenecks through employee-led automation.