Zero-shot prompting is asking an AI model to do a task without giving it any worked examples first.
It is the simplest way to use a large language model: you describe the task in plain language and let the model infer the pattern from its pretraining. That makes it useful when you want:
In practice, many teams start with zero-shot prompting before adding examples or tools.
The model has already learned broad language patterns during pretraining. A zero-shot prompt asks it to apply that prior knowledge to a new instruction without showing input/output pairs.
A typical zero-shot prompt includes:
The model then generates an answer based only on the instruction and the input. This is different from few-shot prompting, where you include examples to steer the output format or behavior.
Zero-shot prompting works best when the task is already common in the model’s training distribution, or when the instruction is very clear.
Prompt:
Classify this email as
billing,technical, orother:
"My invoice was charged twice this month."
Possible response:
billing
No examples were provided; the model inferred the task from the instruction alone.
If you need stable output across many inputs, zero-shot is often a starting point, not the final prompt design.
