OpenAI Academy now has a dedicated course on workspace agents - the automations that handle repeatable tasks inside ChatGPT for Teams and Enterprise without manual prompting each time.
The course, published on OpenAI's Academy site, covers the full setup process: writing agent instructions, connecting tools like web search and code execution, setting trigger conditions, and rolling agents out across a team. The practical framing is aimed at team leads and operations staff who've been running ChatGPT manually and want to shift routine work - weekly reports, meeting summaries, customer request routing - onto a more automated footing.
Workspace agents in ChatGPT work by letting you define what an agent should do, what tools it can access, and when it should run. Once configured, the agent handles those tasks without someone having to type the same prompt over and over.
For anyone already paying for ChatGPT Teams or Enterprise, reading through the Academy material before building anything is worth the time. OpenAI's own documentation reflects how they intend agents to be structured, which usually means fewer compatibility issues and less debugging than building from scratch based on general advice.