ChatGPT agent books a movie and coordinates childcare
Schwartz reports ChatGPT agent buying two movie tickets after payment takeover and confirmation, then creating movie and babysitting calendar invitations.
techradar.com
Original prompt
Shared by the source author. See source notes for context.
Help me plan a movie date night to see Superman with my wife Saturday evening.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- Eric Hal Schwartz on TechRadar
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Schwartz reports ChatGPT agent buying two movie tickets after payment takeover and confirmation, then creating movie and babysitting calendar invitations.
- Human involvement
- Schwartz chooses seats and showing, enters payment/account details himself, confirms the ticket purchase and agrees to the babysitting event and invitees.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Specify the movie, preferred theatre and childcare coordination request.
- Clarify the showing and seats with the agent.
- Let the agent navigate to checkout, then take over payment details.
Eric Hal Schwartz tested ChatGPT agent by asking it to organise a movie date night with his wife. The request named a preferred theatre and movie, included childcare timing, and asked for a more personal invitation. The agent first clarified the showing and seating preferences rather than choosing all the details silently.
It opened the theatre’s website, selected two seats and stopped at payment. Schwartz entered his account details himself, then confirmed the purchase again. He reports that the tickets were bought. With calendar access, the agent created a movie event, asked for invitees and sent invitations; it also made a separate babysitting event after obtaining his agreement. A Slides-based invitation graphic carried the evening’s details in a more playful form.
The copyable workflow is to combine a reservation with the surrounding coordination, while treating credentials, payment and external invitations as separate checkpoints. This was one writer’s successful test, not a guarantee that every booking website will work or that calendar details will always be correct. The account also makes clear that the user supplied preferences and approved meaningful actions. Its reported outcome is a booked evening with coordinated events, rather than unattended control of the household calendar.
Prompt
Original prompt
Source: Blog · Eric Hal Schwartz on TechRadar
Frequently asked questions
Who entered the payment details?
Schwartz says the agent paused at checkout and he entered his account details himself before confirming the purchase.
How was childcare coordinated?
The agent asked whether to create a separate babysitting event, then created it with accompanying invitations after he agreed.
Is this a universal booking guarantee?
No. The source is one successful firsthand test at the writer’s chosen theatre.
More details
Remaining steps
- Confirm the purchase and supply calendar invitees.
- Approve a separate babysitting event and review the invitation graphic.
About this account
The agent navigates the theatre, selects seats, prepares checkout, creates calendar events and makes a Slides invitation around the user’s approved plan.
Firsthand account by Eric Hal Schwartz on TechRadar; original publication date verified against Primary page publication date. Outcomes are the author’s reports, not independently reproduced. Original prompt excerpt transcribed from the primary account; excerpt only, not the complete prompt. Preserve the source-specific context when adapting it.
Credit: Eric Hal Schwartz on TechRadar
Source published
Added to Agents In Use
Source checked
Tools and topics
ChatGPT agent browser · Calendar · Slides
Movie booking · Calendar invitations · Childcare coordination
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.