ChatGPT agent to test hotel date pickers
Tamsin Deasey-Weinstein tested ChatGPT agent on hotel research with a compact brief: location, dates, travellers, budget and preferences for a modern hotel near transport. She wanted to see how the new agent behaved during an ordinary travel-shopping task.
linkedin.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Research modern hotels near public transport for these exact dates, travellers and budget. Show which sites and filters you used, verify the selected calendar dates, and compare only date-specific availability. Stop for review if the picker fails or further research produces no new options; do not book.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- Tamsin Deasey-Weinstein
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Four hotel options, a failed date-picker interaction and no new choices after extended research; no booking is reported.
- Human involvement
- Tamsin observed the browser run and interrupted to question site choices.
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
- Supply location, dates, travellers, budget and hotel preferences.
- Observe filtering and ask about selected search sites.
- Check the actual dates and whether additional research changes the shortlist.
Tamsin Deasey-Weinstein tested ChatGPT agent on hotel research with a compact brief: location, dates, travellers, budget and preferences for a modern hotel near transport. She wanted to see how the new agent behaved during an ordinary travel-shopping task.
Some parts worked. When the available options were numerous, the agent found filters to narrow them. It also answered an interruption about why it had selected particular websites, then continued the work. The author found the explanation more considered than simply choosing the first search result.
The browser interaction was less reliable. Selecting the dates numbered 26 and 29 took minutes. The agent described positions in the calendar but still struggled to make the selection. More time also failed to broaden the research: after twenty-four minutes, it returned the same four choices it had shown after roughly two minutes, having searched only two websites.
The account does not report a hotel reservation or a completed trip. Its useful outcome is a bounded evaluation of the search process: filtering and explaining choices worked better than manipulating a calendar or expanding coverage. A reader copying this task should distinguish a shortlist from a verified date-specific booking result.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: www.linkedin.com · Tamsin Deasey-Weinstein
Frequently asked questions
What worked during the search?
Filters narrowed abundant options, and the agent answered questions about website selection.
What failed in the browser?
It struggled for minutes to select dates numbered 26 and 29.
Did longer research improve the shortlist?
The same four options remained after twenty-four minutes as after about two minutes.
More details
About this account
The agent filtered hotel options and researched two websites, but struggled with calendar selection.
Firsthand primary post and original publication metadata opened. The source summarizes the brief rather than supplying a full verbatim instruction. Suggested prompt written by Agents In Use from that described task.
Credit: Tamsin Deasey-Weinstein
Source published
Added to Agents In Use
Source checked
Tools and topics
ChatGPT agent · Hotel websites
Hotel research · Calendar picker · Search coverage
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.