Using AI Teams in Codex
Codex can serve as an execution environment for AI Team work, particularly when a job benefits from access to files, repositories, coding tools, repeatable tasks, or longer-running execution.
The AI Team remains the workforce definition. Codex is one runtime through which selected Digital Employees, skills, and tools can execute.
Where Codex fits
A representative path is:
Human direction → AI Team capability → Digital Employee + skill → Codex execution environment → Approved files/tools/services → Structured work product
Examples can include:
- Working against an approved repository or document set
- Producing or transforming structured artifacts
- Running repeatable analysis or operational routines
- Using approved tools or MCP services exposed to the execution environment
- Performing scheduled work when the task has a defined cadence and output
Plugin and runtime separation
Accessing an AI Team capability through a plugin does not make the plugin the canonical workforce definition. The plugin packages selected capabilities for the execution environment while the job, skill, workflow, governance, and customer configuration remain separate architectural concerns.
This allows the same AI Team to execute through other supported environments without redefining the business job.
For current Codex capabilities and setup, use the official OpenAI Codex documentation.
See AI Team Execution Layer and AI Agent Tools and Enterprise Integrations.