One-shot prompting is the prompting strategy that provides exactly one demonstration example before the target task - it offers a lightweight way to steer output format and behavior when context is limited.

What Is One-shot prompting?

• Definition: Single-example in-context prompt that illustrates desired mapping or response structure.
• Use Objective: Give enough guidance to reduce ambiguity while minimizing token overhead.
• Common Scenario: Structured outputs such as JSON, classification labels, or templated summaries.
• Performance Profile: Usually stronger than zero-shot for format adherence, but less robust than few-shot on complex tasks.

Why One-shot prompting Matters

• Token Efficiency: Delivers meaningful steering with minimal prompt length increase.
• Format Reliability: A single concrete example often improves schema compliance significantly.
• Fast Iteration: Easy to update and test during application development.
• Cost Control: Lower context use helps manage latency and inference cost at scale.
• Operational Simplicity: Useful default when full few-shot context is unavailable.

How It Is Used in Practice

• Example Selection: Choose a representative example with clear structure and no ambiguity.
• Instruction Pairing: Combine concise rules with the one-shot demonstration.
• Validation Checks: Test against edge cases to confirm the single example generalizes adequately.

One-shot prompting is an efficient middle ground between zero-shot and few-shot prompting - it provides targeted guidance with low token cost and strong practical utility in production systems.