OpenAI adds persistent cloud tasks to ChatGPT for team workflows

The critical detail buried in the announcement is that these tasks run persistently in the cloud with shared oversight, not just on individual machines. This shifts ChatGPT from a tool you invoke to infrastructure that operates independently between human interventions, which changes the liability and governance model for teams deploying it.

This is the logical third layer in OpenAI's enterprise embedding strategy. Three days ago, Space introduced shared conversation history and project context as collaboration infrastructure. Today's persistent task automation extends that logic: if ChatGPT lives in team workspaces with shared context, it should also handle recurring processes without constant human triggering. The pattern mirrors what we saw with Codex in code review workflows (October 12) and Block's security automation (October 13 and 14), where LLMs shift from interactive tools to background processes that compress decision cycles. The difference here is scope: legal intake, contract review, and similar document workflows affect non-technical teams at scale, not just developers or security analysts.

If OpenAI ships audit logging and approval gates for these persistent tasks within 60 days, it signals they're treating this as a governance-first enterprise feature. If adoption stalls without those controls, it suggests teams aren't ready to trust unsupervised LLM execution on sensitive workflows, which would undercut the entire positioning.