实测 Teamily AI 2.0 的 Loop 多助理写稿流程
作者用 3 个助理加自己跑 AI 新闻稿：找选题、核原文、写稿，助理互相纠错但最终仍由人拍板。
I tested Loop in @Teamily_AI 2.0 on my AI news post. 3 agents and me. News Finder finds the story, Researcher checks it on the original page, Writer drafts the LinkedIn post, then it comes to me. one @ on News Finder and it ran on its own. it picked OpenAI's dots launch. the Researcher checked it on OpenAI's own page and left a note for the Writer: "do NOT say unlimited". honestly the Writer still wrote "skips usage limits". so I told it to use OpenAI's exact wording and cut a couplet. it fixed both and I approved it. one chatbot doesn't push back on itself like that. what stood out wasn't just the automation. it's a team workflow I can run again. the agents check each other, share the context (I didn't paste anything between them), and the final call still comes back to me. try it yourself, link in the reply.