Plan a trip, then tighten the rules after feedback

A travel agent found a flight-and-hotel option using stored preferences, then changed its memory after a debrief and searched again.

Done with Hermes Agentยท Intermediate

What they did
A developer set up Hermes Agent with saved traveler profiles, memory, skills, and web search, then asked it to plan a Dubai business trip. It ran three web searches, used the preferences already in memory, and returned one trip option for approval. After the author wrote a debrief, the agent updated its memory file with a new layover limit and replanned.
How it went
The first plan came back at about $850 and stayed pending approval, so nothing was purchased. After the debrief, the agent switched to nonstop flights on the next pass. The profiles and dates were fixture data, so no real trip was booked.
Tips
Load real traveler profiles into memory before trying this with your own trips. Keep the approval step in place so the agent can research and compare without placing orders or making reservations.

Try it yourself with Hermes Agent

Help me plan a business trip from [departure city] to [destination] for [start date] through [end date]. Use my saved preferences for [airline], [seat], [hotel chain], [budget], and [area]. Search flights and hotels, then give me one recommended itinerary with a cost breakdown and one backup option. Stop after drafting the plan and ask for my approval before booking anything. After the trip, ask me for a short debrief, update your memory with any new travel rules, and use those updated preferences next time.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on dev.to โ†—

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.