1. Define success
Name the decision, relevant context, constraints, and evaluation criteria before asking the model to answer.
FREE STARTER SPRINT
Start with prompt quality: make the goal, context, and constraints clear. Compare a before-and-after prompt, inspect an illustrative output, then try three practice decisions. Lessons, results, and explanations are free without an account.
WHAT THIS FREE SPRINT TRAINS
Prompt engineering improves when you can connect a technique to a specific failure mode. This sprint teaches three high-leverage patterns, then asks you to choose the stronger response in unfamiliar situations. The aim is not to memorise a template; it is to recognise when a clearer instruction, a representative example, or retrieved evidence changes the reliability of the result.
Name the decision, relevant context, constraints, and evaluation criteria before asking the model to answer.
Use examples when structure must repeat and grounding when changing or private facts must be supported.
Answer fresh decisions, inspect the explanation, and carry the weak pattern into a personal practice route.
Need the underlying concepts first? Read the prompt engineering guide. Want to find your starting gap across prompting, examples, and grounding? Take the free prompt skills check.
01 · PROMPT QUALITY
A dependable prompt names the goal, the relevant context, the constraints, and how a strong answer will be judged.
Fictional teaching example. The output is illustrative, not a captured Claude response or a reliability guarantee.
Meeting notes: Deniz will prepare the proposal. No deadline was agreed. The team approved a pilot with two customers.
Summarize the meeting.
From these notes, list decisions and actions separately. For each action, include the owner and deadline. Write ‘not specified’ for missing details. Do not add facts that are absent from the notes.
Decision: Run a pilot with two customers. Action: Prepare the proposal. Owner: Deniz. Deadline: not specified.
Next, use the pattern: In the practice round, apply this habit to a quote: make assumptions and missing inputs visible before relying on a total.