Manus to test hotel checkout handoffs
Genevieve Horchler tested Manus against 11 hotel booking engines using the same room request and September dates. The task was to navigate hotel websites, select the stay and room, and see how far the agent could progress through different booking flows.
hotelnewsresource.com
Original prompt
Shared by the source author. See source notes for context.
Please book me a room at [hotel name] from September 15th to 17th.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- Genevieve Horchler
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Eight of 11 tested engines reached guest/payment details; date selectors and CAPTCHAs caused failures or delays.
- Human involvement
- The tester supplied hotel choices; guest and payment details remained a human handoff rather than a confirmed paid reservation.
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
- Give Manus a named hotel and explicit check-in and check-out dates.
- Observe room selection and verify the dates after any repeated calendar attempts.
- Take over at guest and payment details, noting any change from direct booking to an OTA.
Genevieve Horchler tested Manus against 11 hotel booking engines using the same room request and September dates. The task was to navigate hotel websites, select the stay and room, and see how far the agent could progress through different booking flows.
Eight tests were counted as successful, including several that required extra attempts. That label had a narrow meaning: Manus reached the page for guest and payment details. It did not mean a paid reservation or a confirmation page. The traveler still needed to provide personal information or take over the browser.
Date controls were a recurring obstacle. Some engines showed familiar month calendars, while others used a length-of-stay design that Manus could not navigate reliably. CAPTCHAs blocked other attempts. When a direct hotel site failed, the agent repeatedly moved toward Booking.com as an alternative channel.
The account also tested a less specific San Diego hotel request. Manus used web search, surfaced options and continued after the author chose a hotel. The practical outcome is a recorded boundary between navigating availability and completing a booking. A reader can copy the clear dates and hotel choice, then check both the selected dates and the channel used before entering guest information or paying.
Prompt
Original prompt
Source: www.hotelnewsresource.com · Genevieve Horchler
Frequently asked questions
What did eight successful tests actually mean?
The agent reached guest and payment details, not a paid reservation or confirmation page.
Which controls caused trouble?
Unfamiliar date selectors, especially length-of-stay layouts, and CAPTCHAs caused failures or delays.
What happened when a direct site failed?
Manus repeatedly turned to Booking.com as an alternative booking channel.
More details
About this account
Manus navigated hotel booking engines, selected dates and rooms, and sometimes changed to an OTA after an obstacle.
Firsthand source opened and read. Prompt is a verbatim excerpt of the published original instruction.
Credit: Genevieve Horchler
Source published
Added to Agents In Use
Source checked
Tools and topics
Manus · Hotel booking engines · Booking.com
Hotel checkout · Date selectors · Guest handoff
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.