OpenAI says Oracle has 130,000 active ChatGPT users in new case study

Oracle has 130,000 active ChatGPT users and more than 95,000 active Codex users, according to OpenAI’s new customer case study published October 8, 2026. Oracle executives describe faster recruiting, analysis and incident response—not simply more chat conversations. Those usage figures and time savings are company-reported claims, not independently verified findings.

Giving business questions a company-specific map

Oracle Applications Lab built an ontology—a map of the company’s objects, relationships and rules—to help Codex translate everyday business questions into SQL, the language used to query databases. A user describes the desired outcome. Codex selects internal systems, gathers information and returns an analysis, report or application.

Richard Lam, group vice president of Oracle Applications Lab, described one user receiving an almost immediate answer to a question that usually took hours. She checked it against the manual process and found matching numbers. That is a specific example of a checked result, rather than a broad accuracy measurement.

In production engineering, Codex gathers incident context and surfaces the relevant response playbook for site reliability engineers, who keep systems running. Lam says a typical simple incident that previously took an hour can now be resolved in minutes, leaving engineers more time to guide decisions instead of searching for information.

A repeatable briefing before the hiring conversation

Oracle’s talent acquisition team built a market intelligence tool with ChatGPT Work. Starting with a job description, it researches comparable roles, benchmarks compensation and assesses available talent across relevant locations. The output prepares recruiters for their first discussion with a hiring manager.

Jan Ackerman, Oracle’s senior vice president and global head of talent acquisition, puts preparation at about 15–20 minutes, replacing research that took two to four days. The case study also credits the tool with making intake more consistent: hiring managers receive the same quality of information regardless of which recruiter handles the search.

Faster work does not mean unrestricted access

Oracle’s October 5 security account explains the boundaries behind its deployment. For Enterprise ChatGPT, Oracle built connections to internal systems that preserve employee permissions. Employees approve requested access, and backend systems check each operation. Permission to read information does not authorize sending messages or changing records.

Controls for Codex and operations agents

• Temporary access: production privileges covered by the rollout are granted for approved work and a limited period.
• Human approval: protected sensitive actions cannot proceed without the required sign-off.
• Narrower permissions: access reviews remove authority beyond what the assigned work requires.

Lam also warns that people must work alongside Codex to keep its code maintainable. Oracle’s account pairs faster execution with continued responsibility for architecture, security and the resulting output.