# Optimize prompts

Tune prompt structure, few-shot examples, and token budgets for consistency.

- Date: Jul 14, 2025
- Reading time: 14 min
- Level: Intermediate
- Tags: Optimization, Prompting, Responses

## Takeaways
- Start with a minimal contract and expand only as needed.
- Use few-shot examples to lock format and tone.
- Budget tokens by section and enforce caps.

## Establish a signal hierarchy

Put the most important instruction first and avoid competing goals.

Use short, declarative statements instead of long narrative prompts.

## Use few-shot anchors

Add a single high quality example that mirrors the target output. Keep it short and consistent with your schema.

## Enforce token budgets

Measure your average token usage and cap responses to protect latency and cost.

- Set max tokens per section.
- Summarize long inputs before passing them forward.
- Prefer short summaries for high volume workflows.