OpenAI Workspace Agents Build Hour
Build Hour: Workspace agents in ChatGPT is an official OpenAI walkthrough of workspace agents: Codex-powered shared agents for ChatGPT Business, Enterprise, Edu, and Teachers workspaces. Channel: OpenAI. Uploaded: April 29, 2026. Duration: 37:52. Topic tags: workspace agents, ChatGPT, Codex, enterprise agents, tool permissions, Slack workflows, schedules, skills, admin controls.
The video is more useful than a short launch trailer because it shows the workflow shape. OpenAI staff build a meeting-prep agent that checks a calendar, searches Google Drive and the web, uses Gmail for a scheduled summary, applies a reusable skill, and exposes preview runs before the agent is distributed. They then show a software-review style agent working around Slack requests, policies, tools, and routing decisions. The repeated pattern is not one prompt and one answer. It is a persistent team artifact that can gather context, run in the background, use tools, be shared inside a workspace, and be corrected over time.
For Spiralist themes, the strongest signal is delegated organizational memory. A team process becomes an agent with files, skills, app permissions, schedules, channel surfaces, preview traces, versioned behavior, and admin policy around who may build, publish, and use it. That belongs beside the site's Agent Tool Permission Protocol, Agent Audit and Incident Review, ChatGPT, AI Agents, Tool Use and Function Calling, and Agent Log Becomes the Receipt. The video makes a quiet governance shift visible: the assistant is becoming a repeatable workplace actor, not just a conversational interface.
Evidence and limits: this is a primary-source product session from OpenAI, so it is strong evidence for how OpenAI describes workspace agents and weaker evidence for independent reliability. OpenAI's workspace agents announcement describes shared Codex-powered agents that handle complex, long-running workflows while operating within organizational permissions and controls. OpenAI's Workspace Agents Help Center page documents templates, tools, apps, skills, files, custom MCPs, schedules, Slack channels, access settings, version history, analytics, shared connections, and role-based admin controls. NIST's AI Agent Standards Initiative gives independent policy context for why agent identity, authorization, secure operation, interoperability, and evaluation matter as agents become actors inside real workflows.
Uncertainty should stay visible. The Build Hour does not prove that workspace agents are reliable enough for high-stakes work, that teams will configure least privilege correctly, that preview traces will be reviewed with enough care, that shared, personal, or agent-owned connector credentials will remain understandable to ordinary users, or that scheduled background work will preserve human judgment once it becomes routine. It is best read as a clear product signal: enterprise AI is moving from individual chat assistance toward shared, long-running agents embedded in the tools where organizations already coordinate work.