RingCentral rolls out ChatGPT Work and Codex to every employee
OpenAI published a customer story about RingCentral, a business-communications company with nearly three decades of history that now generates more than $2.6 billion in annual revenue and employs thousands of people worldwide. According to the piece, RingCentral gave ChatGPT Work and Codex to every employee, not just engineers, so that anyone at the company could build products and infrastructure regardless of coding background.
The centerpiece of the rollout was an AI-Native Challenge, sponsored by RingCentral's Office of the CEO. Every participant received ChatGPT Work and Codex and was asked to build a complete, end-to-end project, with no mandated workflow or other constraints, covering planning, implementation, testing, documentation, CI/CD, and iteration. The article states that nearly every participant created a working repository, and that thousands of employees, including non-technical staff and even executives, delivered functioning projects. An unnamed engineering leader who led the challenge is quoted saying the clearest lesson was that AI-native development is not about replacing engineers but amplifying them, with humans staying in the loop on product requirements, business context, architectural decisions, and verification.
RingCentral says it applies the same Codex-enabled approach to its own customer-facing AI product line: AI Receptionist (AIR), AI Virtual Assistant (AVA), and AI Conversation Expert (ACE), as part of what it calls its Agentic Voice AI portfolio, with the goal of shortening the distance between an idea and a shipped feature.
Beyond engineering, the Program Management Office (PMO) adopted the same pattern. Using ChatGPT Work, the PMO built what the article describes as an operating system for program management, replacing scattered notes and chat history with AI-powered workflows for status tracking, reporting, release governance, and knowledge transfer. One concrete application is automated status reporting: workflows that generate notifications from issues tracked across Jira, Google Sheets, CRM systems, and other sources, so a meeting can start with blockers, owners, and actions already identified rather than open questions about what changed. The article credits this shift with letting the PMO handle more projects with greater accuracy, though it gives no figures for either claim.
Key facts
- RingCentral, a business-communications company with more than $2.6 billion in annual revenue and thousands of employees, gave ChatGPT Work and Codex to every employee, not just engineers.
- The company's Office of the CEO sponsored an AI-Native Challenge: participants built a complete end-to-end project, from planning through CI/CD, with no mandated workflow.
- Nearly every participant produced a working repository, and thousands of employees, including non-technical staff and executives, delivered functioning projects.
- RingCentral applies the same Codex-enabled approach to develop its own AI product line: AI Receptionist (AIR), AI Virtual Assistant (AVA), and AI Conversation Expert (ACE).
- The Program Management Office built ChatGPT Work workflows for status tracking, reporting, and release governance, including automated status updates pulled from Jira, Google Sheets, and CRM systems.
Why it matters
This is a vendor-published case study of what happens when a large, established company hands ChatGPT Work and Codex to its entire workforce rather than confining AI tools to engineering. RingCentral frames it as a shift in how the company operates: engineering builds faster, and a back-office function like the PMO restructures its workflow around the same tools. For OpenAI, the piece doubles as evidence that ChatGPT Work and Codex can be adopted company-wide, across both technical and non-technical roles.
Who it affects
RingCentral's own workforce, spanning engineers, non-technical staff, and executives who took part in the AI-Native Challenge, and separately the Program Management Office, which rebuilt its reporting and governance workflows on ChatGPT Work. More broadly, the story is aimed at OpenAI's enterprise customers and prospects evaluating whether to roll out ChatGPT Work and Codex beyond engineering teams.
How to use it
The AI-Native Challenge model described here has no mandated workflow: employees are given the tools and asked to ship a complete project end to end, covering planning, implementation, testing, documentation, and CI/CD. Separately, the PMO's automated status reporting pulls updates from Jira, Google Sheets, and CRM systems into ChatGPT Work workflows for status tracking, reporting, release governance, and knowledge transfer, replacing scattered notes and chat history.
How solid is it
This is a customer story published directly on OpenAI's own blog, not independent reporting, so it reads as promotional material for ChatGPT Work and Codex. The claims about speed, accuracy, and adoption are qualitative: the article does not give a total headcount for the challenge, a participation rate, a time-savings figure, or any measured accuracy improvement, only that revenue is 'more than $2.6 billion' and the workforce is 'thousands' strong. All three passages quoted in the piece, the opening line about putting real AI tools in everyone's hands, the engineering leader who led the challenge, and a PMO team member, are unnamed.
Risks and caveats
Because the piece is OpenAI's own customer marketing, its claims of amplified engineers, greater accuracy, and a company-wide culture shift come from RingCentral and OpenAI rather than a third party, and none of the productivity or accuracy gains are quantified. Readers should treat the framing as a case study built to promote adoption of ChatGPT Work and Codex rather than as an audited outcome.
“AI accelerates the entire development cycle, while humans remain in the loop, guiding product requirements, providing business context, making architectural decisions, and ensuring every outcome is tested and verified.”
— Engineering leader at RingCentral who led the AI-Native Challenge (unnamed in the source)