ASSESSMENT · 60 MIN

AI Operator Test

It doesn't measure what you already know. It measures judgment, autonomy, and your ability to deliver something real in limited time. One hour, one task, one deliverable.

60
minutes, from start to delivery
1
real deliverable, not a quiz
0
prior technical requirements
THE CHALLENGE
Timer: 60 minutes

Take a repetitive task from your work and turn it into a small documented system using Claude.

Replying to one type of email the same way every time, organizing files, summarizing meetings, generating reports, sorting tasks, writing product descriptions. Whatever you repeat and want to stop doing by hand.

No tiene que ser perfecto. Tiene que funcionar, estar documentado y que otra persona pueda usarlo sin preguntarte cómo.

WHAT YOU DELIVER

When you finish, a single Markdown (or plain text) document with five parts.

  1. 01

    The problem

    The repetitive task you chose to solve, in one sentence.

  2. 02

    The prompt

    The exact prompt you gave Claude. No window dressing.

  3. 03

    The result

    The generated output: text, file, or documented process.

  4. 04

    The process

    The checklist of steps you followed, in order.

  5. 05

    The second version

    What you would change or improve if you did it again.

Delivery template 
# AI Operator Test — [Your name]

## 1. Problem
[The repetitive task you chose, in one sentence]

## 2. Prompt used
[The exact prompt you gave Claude]

## 3. Result
[The generated output or the documented process]

## 4. Process
- [ ] Step 1
- [ ] Step 2
- [ ] ...

## 5. Second version
[What you would change if you did it again]
HOW IT'S EVALUATED

There's no numerical score. The full pattern of the deliverable is read.

Clarity What you deliver is understood without further explanation.
Autonomy You decided without waiting to be told how to do it.
Order The deliverable has a logical structure, not a text dump.
Documentation You left a record of what you did and how, not just the result.
Smart use of AI You used AI as a tool, not a crutch. You reviewed the output.
Honesty When you don't know something, you say so. You don't pad it with empty text.
Independence You tried to solve it before asking. You asked once, well.
SIGNALS WE WATCH FOR

IN FAVOR

  • Sums up the problem in one line without being asked.
  • Delivers a clean document, no noise.
  • Asks few questions, but the ones they ask are precise.
  • Detects on their own whether the AI's output is wrong or incomplete.
  • Proposes an improvement at the end without being asked.
  • Structures the work before executing it.

AGAINST

  • Doesn't read the full instructions before starting.
  • Delivers a chaos of notes with no structure.
  • Pastes the AI's output without reviewing or adapting it.
  • Documents the result, but not how they got there.
  • Freezes at the first ambiguity instead of deciding.
  • Asks for part of the work to be done for them.
  • Spends the time asking instead of executing.

If you're unsure at any point, say what you did and why you made that decision.

That's judgment too. What's being assessed isn't whether you know everything, but how you think when you don't.

NEXT STEP

When you have your deliverable, save it. Where you go depends on where you are:

01

Start Level 1

If you're not a Monkey yet, start here. This deliverable counts as initial practice.

→ 02

Collaboration process

If you want to collaborate in the MOONKEY XN ecosystem, here's the onboarding.

→