Our Take
A feature that sounds simple but signals OpenAI is betting on ChatGPT as an ambient work tool, not just a chat interface you initiate.
Why it matters
Scheduled Tasks moves ChatGPT from reactive (you ask, it responds) to proactive (it runs and reports). For teams relying on ChatGPT for routine analysis, monitoring, or reporting, this removes the manual trigger step and surfaces results on a cadence.
Do this week
Product teams: audit your current ChatGPT workflows this week to identify repetitive tasks that run on fixed schedules (daily reports, weekly summaries, periodic checks), then pilot Scheduled Tasks on one low-risk workflow before rolling out to your team.
ChatGPT can now run tasks on a timer
OpenAI has added Scheduled Tasks to ChatGPT, a feature that lets users set jobs to run at specified times without manual initiation. The feature is available now in ChatGPT (company-reported).
Scheduled Tasks allows you to define a task, set a frequency (daily, weekly, or custom), and ChatGPT executes it on that cadence. Results appear in your chat interface or can be delivered via other channels. The feature works with custom GPTs and ChatGPT's native capabilities.
Proactive work changes the utility model
Until now, ChatGPT was strictly reactive: you opened it, prompted it, got a response. Scheduled Tasks inverts that flow. You define a job once, set it to run, and ChatGPT handles the execution and delivery on its own schedule.
For practitioners, this is meaningful in specific domains. Research teams can schedule daily briefing summaries. Support teams can set up periodic ticket reviews. Sales ops can run weekly pipeline analysis. Finance teams can automate routine data pulls and reports. None of these require human intervention on each cycle.
The catch is reliability. Scheduled execution introduces new failure modes: a task misses its window, a result never arrives, an output format degrades. Traditional automation tools (Zapier, Make, cron jobs) have mature monitoring and retry logic. ChatGPT's robustness at scale here is untested in production.
Start with low-stakes, high-repetition tasks
If your workflow involves a task you run more than twice a week on a fixed schedule, and the cost of a missed execution is low, Scheduled Tasks is worth a pilot. Set up one task, monitor it for two weeks, and only then expand to higher-stakes workloads.
Do not schedule mission-critical reporting or decisions on this yet. ChatGPT's SLA and error-handling are not published, and there is no public status page or incident history for Scheduled Tasks availability. Pair it with a fallback or notification system so you know when a task fails.
For teams already embedding ChatGPT into workflows via the API, Scheduled Tasks offers a no-code alternative for simple cadences. If your use case is complex (conditional logic, multi-step dependencies, external service calls), keep using your existing orchestration layer.