> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gc.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Common Prompting Pitfalls

> Common mistakes to avoid when prompting legal AI

## Common Mistakes

### Terminology Issues

* Using ambiguous terms
* Not specifying legal terms of art
* Unclear abbreviations
* Undefined industry jargon

<Warning>
  Always clarify specialized legal terms and ensure the AI understands specific legal context.
</Warning>

### Context Problems

<AccordionGroup>
  <Accordion title="❌ Bad Example">
    Is this okay under Revlon?
  </Accordion>

  <Accordion title="✅ Good Example">
    Analyze whether this board decision meets Revlon duties under Delaware corporate law.
  </Accordion>
</AccordionGroup>

## Structural Issues

### Prompt Construction

* Too vague or general
* Missing critical context
* Unclear objectives
* Information overload

### Common Format Mistakes

```markdown theme={null}
❌ Multiple questions in one prompt
❌ Unclear priority order
❌ Missing output format specification
❌ Ambiguous instructions
```

## Best Practices to Avoid Pitfalls

### Use Easy Prompt

[Easy Prompt™](/guides/prompting/using-easy-prompt) can help you avoid many common pitfalls by automatically refining your prompt before sending. It adds structure, specificity, and context to your questions.

### Use the Skill Library

The [Skill Library](/guides/prompting/using-skills-in-chat) provides ready-to-use templates for common legal tasks. Type `/` in the chat box to attach one, or browse My Skills, Shared with Me, Organization Skills, and GC AI Skills. Starting from a proven template avoids many structural issues.

### Clear Communication

1. Define terms clearly
2. Specify jurisdiction
3. Provide relevant context
4. State desired output format

<Tip>
  When in doubt, break complex questions into smaller, clearer prompts. You can also use the message queue to send follow-up questions while GC AI is still responding.
</Tip>

### Quality Control

* Verify AI understanding
* Cross-check responses
* Document assumptions
* Track successful approaches