Página 1 de 4

Test Reviewers - Travel Project

In this role you will review pre-generated travel tasks. Each task is composed of distinct components that must be evaluated separately:

Persona — A deep description of the person that wants to travel

Intent — what the traveler is trying to achieve

Scenario — the real-world situation in which that intent must be fulfilled

Hidden Context — unstated constraints or risks

Initial Prompt — how the user actually asks for help

Trap Authenticity - a logical flaw or critical dependency (like weather, visas, hours) the AI must spot.

Guidelines for reviewers here

Your job is to determine whether these pieces work together to create a realistic, multi-turn collaboration problem. The goal is to create scenarios and prompts that will require multiple interactions with the AI.

For the last part of this test, you are required to demonstrate a real, multi-turn interaction with an AI assistant.

Please follow the instructions below carefully:

- The conversation must be in English.

- Choose an AI assistant (GPT, Gemini, Claude, or xAI).

- Start with an initial prompt, then continue the conversation over multiple turns, refining or adjusting your request as the interaction evolves.

- The goal is to show iteration and collaboration, not a single one-off question.

What to submit: Paste the full conversation transcript, or

Provide a shareable link to the conversation.

This interaction should reflect how a real person would work with an AI to clarify goals, add constraints, and progressively improve the outcome.