We’re recruiting Finance professionals (financial planners, investment advisors/consultants, and bank relationship managers) to join a project as scenario & prompt reviewers.

Your role in this selection form is to complete a screening review of pre-generated financial scenarios and tasks, assessing whether they feel realistic, believable, and grounded in real-world client situations.

👤  Who can apply
    •    Bank relationship managers
    •    Investment advisors / consultants
    •    Financial planners
    •    3+ years of professional experience

🧩 Selection

In this application, you will review one test scenario/task following the instructions below.
If selected, you will be invited to review 5 real tasks. 

Applications close on: 08/02/2026

Instructions:

Please read before starting!

Part 1 - For this part, you are strictly forbidden from using any AI tools (including ChatGPT, Gemini, Claude, or similar).

In this task, your role is to review pre-generated scenarios and prompts.Your objective is to evaluate whether they represent real people, real situations, and the kinds of questions people genuinely ask for help with in everyday life. You are not judging the quality of any advice, response, or solution.Instead, you are evaluating whether the scenario itself feels authentic and believable — the kind of situation a person would realistically bring to an AI or a professional advisor. The goal of this task is to determine whether the personas, contexts, and prompts are grounded in lived experience, rather than feeling artificial, academic, or “test-like.

What you are rating: Each task is composed of distinct components. You must evaluate each component separately, using your real-world judgment: Persona — A deep description of the person that wants financial advice Does this sound like a real human being? Do the details feel plausible and internally consistent? Intent — What the person is trying to achieve
Is the goal something a real person would have? Is it framed in a realistic, human way? Scenario — The real-world context
Does the situation reflect how life actually works (constraints, uncertainty, trade-offs), or does it feel staged or incomplete in an artificial way? Hidden Context — Unstated constraints or risks
Are there believable, real-world factors that are not explicitly stated but would naturally exist in this situation? Initial Prompt — How the person actually asks for help
Does the wording sound like how people really talk and ask questions, including being vague, incomplete, or imprecise? Trap Authenticity — A logical flaw or critical dependency
Is there a real-world dependency or assumption that could realistically cause failure if overlooked (e.g. weather, visas, taxes, liquidity, timing)?A scenario can be realistic and still poorly constructed — and that is valid to score low. How to score (1–5)Use the following scale for each component

• 5 — Highly authentic
• Feels like a real person in a real situation you would expect to encounter in real life.
• 4 — Mostly realistic
  Plausible and coherent, but slightly simplified or under-specified.
• 3 — Real but shallow
  Common in real life, but low signal or limited realism depth.
• 2 — Weak realism
  Technically possible, but lacks human texture or feels incomplete.
• 1 — Synthetic / artificial
  Feels staged, unnatural, contradictory, or unlike real human behavior.

Expert rationale (required) After scoring, you must write a short expert rationale explaining:

• Why you believe the scenario is realistic or unrealistic, and
• Which elements feel grounded in real life versus artificial or missing.

Focus on: plausibility, internal consistency.

Part 2- Final task — Real multi-turn AI interaction (required)

In this final step, you are required to demonstrate a real interaction with an AI assistant.

⚠️ Do not invent, simulate, or script a conversation.
We are explicitly looking for a true conversation that actually took place.

Instructions:

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

• Start with an initial prompt written by you.

• Continue the interaction over multiple turns, refining or adjusting your request as the conversation evolves naturally.

• The conversation must be in English.

What to submit:

• Paste the full, unedited conversation transcript, or

• Provide a shareable link to the real conversation.

Important notes:

• Single-turn questions are not sufficient. ( We need more than 1 interaction with the model)

• Pre-written or fictional conversations will be disqualified.

• The goal is to observe how you collaborate with an AI: clarifying goals, adding constraints, reacting to outputs, and iterating.

This task is meant to reflect how a real person actually works with AI in practice — step by step, with genuine back-and-forth.