How to use Gemini for a changing family trip plan
Gemini Spark revised a Hershey family plan using connected context and grandparents’ availability, shared the document, and stopped before an Airbnb purchase.
theverge.com
Original prompt
Shared by the source author. See source notes for context.
I’m going to be in Hershey PA with my wife, two kids, and dog the weekend of July 18th.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- David Pierce
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Spark produced and revised a family-trip document and sent a requested email; the Airbnb transaction was not completed.
- Human involvement
- Pierce reviewed personal assumptions, corrected the schedule, changed family constraints and explicitly requested document sharing; booking remained manual.
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
- Pierce supplied the Hershey weekend party and reviewed a Google document assembled from connected context.
- He corrected an assumed nap time and added grandparents, prompting an accommodation revision.
- Spark shared the requested plan by email but could not complete the Airbnb booking.
David Pierce tested Gemini Spark on a weekend trip to Hershey, Pennsylvania, with his wife, two children and dog. Spark produced a Google document that used information from his connected Google account, bringing personal context into the planning process.
Some details were useful and others required correction. The plan incorporated the children’s ages and named the dog without Pierce explicitly providing that information in the request. He could only speculate about some underlying sources. It also selected a nap time that needed adjustment, showing that plausible personalization was not the same as confirmed family knowledge.
When Pierce added that the grandparents could babysit, Spark revised the accommodation plan from a hotel toward an Airbnb. He then asked it to share the document with his wife. It located her email address and sent the requested message, although he found the wording overly formal.
The workflow stopped short of lodging purchase. Spark requested website permission but could not complete the Airbnb booking, instead offering options and details for manual action. The demonstrated result is an editable, context-aware plan and a requested email, alongside a visible booking boundary and human correction of assumptions.
Prompt
Original prompt
Source: The Verge · David Pierce
Frequently asked questions
Did Spark book the Airbnb?
No. It encountered a booking boundary and supplied options for manual action.
Was its personal context always accurate?
No. Pierce corrected a nap-time assumption and could not verify the origin of every personal detail.
What changed after grandparents were added?
Their availability for babysitting caused Spark to revise the accommodation plan toward an Airbnb.
More details
About this account
Pierce supplied the Hershey weekend party and reviewed a Google document assembled from connected context. He corrected an assumed nap time and added grandparents, prompting an accommodation revision.
Full original article read through primary HTTP HTML. Original prompt excerpt. The author speculates about some context sources; these are not asserted as proven data provenance. No completed accommodation purchase. Original prompt excerpt.
Credit: David Pierce — The Verge
Source published
Added to Agents In Use
Source checked
Tools and topics
Gemini Spark · Google Docs · Gmail · Airbnb
family travel · personal context · booking boundary
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.