Write strong instructions
Want more background?
The idea in everyday life
Write instructions like a compact work order: objective, audience, source context, boundaries, output form, and acceptance criteria. Examples help when style or categorization is hard to describe. Ask the model to expose uncertainty rather than invent missing facts.
A prompt is the instruction you give an AI tool. Include the goal, background, source material, limits, an example, and how you will check the result.
One step at a time.
Use the sample or information you have permission to use. Keep your work as you go. You can pause between steps.
- 1
State the task and real audience in the first lines instead of assigning a vague expert persona.
- 2
Provide only the source material and background required, clearly separating instructions from reference text.
- 3
Specify must-include facts, prohibited assumptions, desired structure, length, and tone.
- 4
Ask for missing information, uncertainty flags, and a final self-check against explicit criteria.
See how it works.
The manager provides a transcript and says: produce decisions, action items with owner and due date only when explicitly stated, unresolved questions, and a 100-word recap. Quote the timestamp for each action, write 'not specified' for missing owners, and do not infer commitments.
Did you cover these?
- Required facts and supplied facts are clearly separated.
- The output format makes omissions and review easy to detect.
- Instructions state what to do when information is missing or conflicting.
What to finish
A reusable prompt template and one completed brief for a real task.
Watch for this
Adding adjectives such as ‘world-class’ while omitting data, constraints, and a definition of done.
How can I tell I understand it?
Another person could evaluate the output without asking what ‘good’ means.
Quick knowledge check
Answer all 3 questions. Get at least 2 right and the lesson completes automatically.
You’ll get an explanation for each answer. Signed in? Your results save to your account.
Pass the check. We’ll save the progress.
Answer all 3 questions above and get at least 2 right. You can review the explanations and try again.
Go to the knowledge checkOptional course overview video · same video as Lesson 1
AI and Machine Learning with Google Cloud
From Google Cloud. Watch here or open it on YouTube .
Pause when the instructors add context, delimit input, or specify output. Label those parts in your own reusable instruction block.Tools, templates, and lesson sources
Tools and templates
Tools open in a new tab so you can keep your place here.
Sources and further reading
Explore the references behind this lesson.
Report a problem with this lesson
Tell us what went wrong. Please leave out personal information and private project contents.