Skip to content
agentusehub
Sign in+ Submit
Travel & BookingsAny

Plan business trips and learn from each debrief

A demo travel-ops agent searched flights and hotels, applied stored preferences and produced one recommendation pending approval, then updated its memory; nothing was booked.

Read the experience
Time saved
Not reported
Setup time
Not reported
Running cost
Not reported
Cost range
Not reported

Use case overview

A demo travel-ops agent searched flights and hotels, applied stored preferences and produced one recommendation pending approval, then updated its memory; nothing was booked.

Read the original report ↗

Reported experience

  1. What they did

    Built on Hermes Agent's own memory, skills and web search, running on gemini-3-flash-preview. With seeded traveler profiles, it ran three web searches for a Dubai trip and applied stored preferences unprompted. After a debrief typed by the author, it rewrote its memory file with a 3-hour layover cap and re-planned.

  2. How it went

    The first plan was one recommendation (about $850) ending at pending approval. The re-plan enforced non-stop flights because of the debrief lesson. Profiles and trip dates were fixture data, and nothing was booked.

  3. Worth knowing

    Demo data was seeded before recording and trip dates are fixtures, so real use needs your own traveler profiles loaded into memory first.

Example prompts

Try with Hermes Agent
Plan a business trip to [destination] from [date] to [date], departing from [city]. Search flights and hotels, apply my preferences in [seat, airline, hotel chain, budget, neighborhood], and give me one recommendation with a price breakdown and one backup. Do not book anything; wait for my approval, and after the trip ask for a debrief and save what you learned to memory.

Related use cases