Scheduled ChatGPT Tasks Turn Recurring Work Into Structured Workflows
OpenAI’s Danielle Zaghian presents scheduled tasks in ChatGPT as a way to turn recurring work into defined workflows: specify the sources, output, schedule and limits, then have ChatGPT run the task when needed. Her examples range from a read-only weekday brief that prioritizes work across Slack, email and calendar to a weekly feedback triage that ranks issues, assigns owners and drafts team updates. Converting the latter into a Workspace Agent moves it from a laptop-based task into a shared cloud workflow available in ChatGPT or Slack.
The examples are configured with sources, a fixed output, a schedule, and limits
? danielle-zaghian presents scheduled tasks in the new ChatGPT desktop app as a way to prioritize recurring work. A useful pattern in her examples is to specify where the task should look, what decision-oriented result it should produce, when it should run, and what it is permitted to do.
Her “Chief of Staff” example is a personal prioritization task connected to Slack, email, and calendar; its configuration also reviews Notion for onboarding feedback and Google Drive for survey responses and review files. The operating instruction is specific: return a concise brief in a fixed order—Top Line, Do Today, Meetings, Product Feedback Snapshot, Waiting On, and Can Wait. It is told to prioritize actions and deadlines, use read-only access, and neither send messages nor modify content across connected services.
That structure separates immediate decisions from ambient information. The sample brief puts a deadline to review revised onboarding screenshots by 11:00 AM at the top. It also calls for a decision on a Python 3.13 setup fix before a 1:00 PM launch review, identifies meetings and preparation needs, and places an informational finance update under “Can Wait.”
Zaghian schedules the task with the prompt, “Send me this brief every weekday morning at 8.” ChatGPT confirms a recurring automation called “Daily Chief of Staff Brief.” For her, the value is that the prioritization arrives during morning coffee rather than requiring a manual reconstruction of the day from several systems.
The next step is not more summary, but a work product with owners
The daily brief is read-only: it helps one person decide what deserves attention. Zaghian’s second workflow has a different job. Technical-onboarding feedback, she says, arrives through Slack, Notion, and Google Drive; the scheduled “Feedback Triage” task consolidates those signals each week, summarizes them in a Google Doc, and drafts a post for the team Slack channel.
The task runs in a separate “Feedback Triage” chat every Monday at 8:30 AM. Its result does more than collect feedback counts. It ranks issues, connects each finding to a stated signal, and gives a named person a concrete next step.
| Rank | Finding | Signal | Follow-up |
|---|---|---|---|
| 1 | Workspace access is unclear; new hires cannot tell which role unlocks the API quickstart, while two request-instruction versions circulate. | 11 mentions across Slack and the workshop survey | Tessa Rowan: replace the role matrix and quickstart screenshot. |
| 2 | Python 3.13 breaks local setup before the first API request at dependency installation. | 4 Notion reports | Dorian Pike: test the pinned dependency set and add a temporary workaround. |
| 3 | New hires want role-specific paths for engineers and solutions architects. | 7 requests across Notion and the workshop survey | Elara Voss: draft two role-specific learning paths. |
Those fields make the recurring process actionable rather than merely descriptive. A team can see which issue is most prominent, what evidence supports that ranking, who is responsible for a response, and what that response is meant to be. The Slack-post preview reduces the result to a team-level update: workspace-access guidance is the top issue, followed by the Python 3.13 setup failure and requests for role-specific learning paths; the owners and next steps remain in the fuller output.
This workflow has broader outputs and actions than the morning brief. It is configured to create a Google Doc and draft a Slack post, not simply review connected systems and return a prioritized list. Zaghian notes one remaining constraint: in this setup, the weekly triage runs on her laptop.
A shared agent moves the workflow beyond its creator's laptop
That laptop-bound constraint marks the transition from a personal scheduled task to a shared workflow. ? danielle-zaghian describes a Workspace Agent as a way to take a local workflow such as Feedback Triage and run it in the cloud as a separate agent.
The conversion shown is direct. The user asks ChatGPT to turn the Feedback Triage workflow into a Workspace Agent called “Product Feedback Router.” ChatGPT first creates it as a draft and presents a Create control; after creation, the interface marks the agent ready to try in chat.
Now that it's in the cloud, my team can use it directly, either from ChatGPT or in Slack.
The distinction is practical: the Chief of Staff task assembles read-only context for personal prioritization, while Feedback Triage produces a document and a proposed team update. As a Workspace Agent, that triage workflow becomes available to the team through ChatGPT or Slack rather than remaining a local scheduled job.