BuildSpark AI

Map · Foundations · 03 · Prompting

Concept card · Foundations

Prompting

Prompting is how you state the task this turn: the reader, the shape, and what to leave out.

How you state the task this turn: the reader, the shape, and what to leave out.

Last checked 2026-10-16 Free card 4 min

Ask about Prompting

What changed

Oct 16: The card uses the v2 layout. Prerequisites and related concepts come from the registry. The prompting guidance itself is unchanged.

1 What it is

Prompting is how you state the task for this turn: the instruction, the material, and the constraints. A weak prompt and a strong prompt can ask for the same outcome and still give the model different jobs.

Same task: summarize meeting notes for someone who was not there. Toggle the prompt. Highlights show what the strong version adds.

summarize this [paste notes here]
No readerNo output shapeNothing to leave outNo rule if unclear
  • ReaderNot named.Named: a teammate who was not there.
  • Output shapeNot set.Three short sections: decisions, owners, open questions.
  • Leave outNot set.Small talk stays out.
  • If unclearNot set.Say it is unclear. Do not invent a decision.

2 Why it matters

Cost

A vague prompt often leads to another turn to get something you can use.

Quality

Naming the reader, the format, and what to leave out gives the reply a finish line.

Risk

An open prompt is easier to steer off the task, including by text inside a document you attached.

3 How to apply · try today

Rewrite one prompt you use so it names the reader, the output shape, and one thing to leave out.

Common mistake

Adding adjectives (“be detailed and insightful”) instead of a concrete constraint.

Self-check

Which change makes the meeting-notes prompt stronger?

Questions

What is prompting?
Prompting is how you state the task this turn: the reader, the shape, and what to leave out.
What makes the meeting-notes prompt stronger?
Name the reader, the output shape, and one thing to leave out.
What is a common prompting mistake?
Adding adjectives (“be detailed and insightful”) instead of a concrete constraint.