In a case study published November 5, 2026, OpenAI described how cult.fit, an India-based fitness company operating 708 centers across 77 cities and serving nearly one million paid members, has adopted its Codex coding tool. According to the account, cult.fit first standardized its engineering teams on Codex after previously using a mix of coding tools with inconsistent conventions. The company then extended Codex into specialized workflows, including a Site Reliability Engineering agent that investigates incidents and can propose fixes for an engineer to review, and security agents that help identify vulnerabilities. CTO Santhosh Kumar said the company wanted to become 'more AI-native in the way we run things.' OpenAI reports that cult.fit made company data easier for Codex to access, reducing reliance on a central analyst team, and that more than 20 internal applications have been built by non-engineering teams in recent months. HR, for example, built a workflow for performance reviews. The study states that when an internal dashboard tool's cost rose sharply, an engineer rebuilt the functionality over a weekend, avoiding more than $100,000 in annual software costs. Cult.fit also uses the OpenAI API for an AI Trainer feature, now built around a voice-first experience using GPT Realtime. The figures and claims come from OpenAI's own account, not independent verification.
- Case study published by OpenAI on November 5, 2026
- Cult.fit operates 708 fitness centers across 77 cities, nearly one million paid members
- More than 20 internal applications reportedly built by non-engineering teams in recent months
- Rebuilding one internal dashboard tool reportedly avoided $100,000+ in annual software costs
- A social feed estimated at over a year was reportedly built and shipped in two months
What it means for you
This is a vendor telling its own success story, so read the numbers as claims rather than proven results. But there's a genuinely useful idea underneath the marketing: cult.fit didn't just hand everyone an AI tool — it first standardized how its engineers worked, mapped its data so people could actually use it, then ran hackathons and twice-weekly office hours to help teams ship real things. The savings and speed came from that groundwork, not from the tool alone. The one transferable lesson is that AI let non-engineers rebuild or replace software they were paying for — worth knowing if your software bills keep climbing.
Try this
Pull your list of recurring software subscriptions and flag any single tool where the cost jumped this year. Before shopping for an alternative, write down in plain English exactly what that tool does for you — that one page is the starting point for deciding whether it's worth keeping, replacing, or rebuilding.
Who should care
Owners and operations leads at mid-sized organizations with an in-house engineering team, plus whoever signs off on software subscriptions and whoever runs internal data requests through a bottlenecked analyst.
Skip this if
You have no internal developers, your software costs are flat and manageable, or you're not currently feeling friction in how teams get data and small tools built — in which case this is just a case study about someone else's company.
Sources: OpenAI — read the original