AUTA AI Academy crest AUTA AI ACADEMY
Home · Module 2 · Lesson 2
Module 2 · Lesson 2 · Week 2

Advanced Prompting: Examples, Iteration and Repair

Few-shot, critique loops and rescue moves

⬇ Study guide PDF

In this lesson you will

  • Use few-shot examples to control style precisely
  • Run a critique-and-revise loop with the same model
  • Rescue a failing conversation with three repair moves

Few-shot: show, don't tell

Giving two or three input/output examples ('few-shot') pins style and structure better than any adjective. Want minutes in your organisation's format? Paste one old set of minutes as the example.

The critique loop

After a draft, ask: 'List the 3 weakest points of this draft and why, then revise it addressing them.' The model critiques far more reliably than it creates. Two loops usually lift a draft from fine to sharp.

Three repair moves

When output drifts: (1) restate the goal and constraints in full - the window is crowded; (2) ask for an outline before the full text; (3) change the angle - new role or new format. Never argue with the model; re-specify it.

Key terms

Few-shot prompting

Including worked examples inside the prompt.

Critique loop

Model reviews then revises its own draft.

Outline-first

Requesting structure before prose.

Re-specification

Restating goal/constraints to fix drift.

🛠 Hands-on task: Two-loop upgrade

Tools: Any free chatbot

  • Draft a 200-word message with a structured prompt.
  • Run the critique loop twice, keeping each version.
  • Note which loop added the most value and why.
Key takeaways: Before you go, three takeaways from this lesson. One: Use few-shot examples to control style precisely. Two: Run a critique-and-revise loop with the same model. Three: Rescue a failing conversation with three repair moves. And your hands-on task: two-loop upgrade. Draft a 200-word message with a structured prompt. When it is done, mark the lesson complete on your dashboard, keep the artifact for your portfolio, and step into the next lesson. I will see you there.
← 1. Prompting 101: The Anatomy of a Prompt 3. AI for Study & Research: Learn Twice a →