Learn AI by proving it

A skill tree for AI, where every branch is earned.

Short modules you finish at your own pace. Each one ends with a task on your own work. Pass it, unlock the next. Finish the trunk and earn a badge an employer can verify.

Start the trunkSee the tree

How it works

  1. Learn

    A module takes under 30 minutes and uses tools you already have. No installs, no API keys.

  2. Prove

    Finish with a task on your own material, so the result only makes sense as yours.

  3. Unlock

    A grader marks it pass or not yet, with one sentence of feedback. Retries are free.

The trunk

Four modules everyone completes before branching.

  1. 1. What you are actually talking toF1 · F2 · F3 · ~30 min
  2. 2. Sharing safelyF5 · F4 · ~25 min
  3. 3. Writing a prompt that worksP1 · P2 · ~30 min
  4. 4. Deciding what to trustV1 · V3 · ~30 min

Two branches at launch

Knowledge worker

You write, analyze, plan, or manage. You'll use AI through chat and document tools and need to know what to trust.

Branch modules coming after the pilot

Builder

You'll wire AI into a tool, script, or workflow. You need APIs, evaluation, and cost, on top of the fundamentals.

Branch modules coming after the pilot

The skill tree

Green markers are trunk skills. Each skill is a thing you can do, and each badge lists exactly which ones it certifies.

Foundations

  • F1Explain in plain words what a language model does (predicts text from context) and name two things it is NOT (a database, a search engine).
  • F2Identify a hallucination in a model response and explain why it happened.needs F1
  • F3Explain what context is, what fits in it, and why the model "forgets".needs F1
  • F4Recognize which tasks a model is good at, bad at, and unsafe for, given a list of real tasks.needs F1, F2
  • F5Describe what can and cannot be shared with an AI tool (personal, confidential, credentials) and apply it to a real scenario.

Prompting

  • P1Write a prompt with role, context, constraints, and an output format, and show the result improves over a bare question.needs F1, F3
  • P2Iterate on a prompt: diagnose a bad output, change one thing, and explain what the change fixed.needs P1, F2
  • P3Provide examples in a prompt (few-shot) to steer format or tone.needs P1
  • P4Break a task the model fails at into steps it can do, and chain them.needs P2

Verification & judgment

  • V1Check a model's factual claims against a source and mark which held up.needs F2
  • V2Write a short rubric for a task and use it to compare two model outputs.needs P1
  • V3Decide when to trust, verify, or reject a model output based on stakes and reversibility.needs F4, V1

Tools & workflow

  • T1Pick an appropriate tool for a task (chat, document upload, code assistant, image) and justify the choice.needs F4
  • T2Use a model on their own material (a document, a dataset, notes) and produce a usable result.needs P1, F5
  • T3Turn a repeated manual task into a reusable prompt or template.needs P3, T2
  • T4Explain cost and rate limits in plain words and choose a model tier for a task.needs F3, T1

Communication

  • C1Explain to a non-technical person what they did with AI and what they checked.needs F1, V1
  • C2Document a prompt or workflow so someone else can reuse it.needs T3