GlobSynkGlobSynkPrompt Lab™
Foundations · The language of AI

Prompt Engineering

Prompt engineering is the skill of asking AI clearly. A prompt is your instruction — and small changes in how you phrase it can make a big difference in what you get back. It's the first step in learning the language of AI.

Prompt basics

A good prompt usually answers three questions for the AI:

  • Who should it act as? (a teacher, an editor, a marketer)
  • What exactly do you want it to do?
  • How should the answer look? (length, tone, format)

Roles

Giving the AI a role focuses its response. "You are a patient teacher explaining to a beginner" produces a very different answer than "You are a senior analyst." The role sets vocabulary, depth, and tone.

Structure

Structure your request so the AI can follow it. Put the instruction first, then any material to work on, then the desired format. Clear sections beat one long run-on sentence.

Patterns

Role + Task + Format
Tell the AI who to be, what to do, and how to answer.
Few-shot examples
Show 1–3 examples of the output you want, then ask for more.
Step-by-step
Ask the AI to work through a problem in clear steps.
Constraints
State limits: length, tone, audience, what to avoid.
Ask-back
Invite the AI to ask clarifying questions before answering.
Iterate
Refine: 'shorter', 'more concrete', 'change the tone'.

A good example

Instead of:

write about marketing

Try:

You are a friendly marketing coach.
Write 3 social post ideas for a small bakery
opening next week. Audience: local families.
Keep each under 30 words, warm and upbeat.
Next

From prompts to context

Great prompts are only the start. As tasks get bigger, context — what the AI knows about your goal, history, and materials — matters even more.

Learn Context Engineering →