Effective Patterns

Overview

Getting the most out of AI assistants requires understanding some key patterns and best practices. Let's explore some proven patterns that will help you get better results from your AI assistant interactions.

Markdown Formatting

AI assistants respond well to instructions in markdown formatting. For example:

# Instructions

## Role
This is your role, etc...

## Rules
**Always** follow these rules:
- Rule 1
- Rule 2
- Rule 3

## Report Structure
This is the structure of the report:
{
  "title": String,
  "topics": Array[String],
  "language": String,
  "description": String
}

## Report Output Example
This is an example of the report output:
{
  "title": "This is the title of the report",
  "topics": ["topic 1", "topic 2"],
  "language": "EN",
  "description": "This is the description of the report"
}

Tags

AI assistants respond well to instructions in tags. For example:

<instructions>
<identity>
This is your identity, etc...
</identity>
<report>
<report_structure>
{
  "title": String,
  "topics": Array[String],
  "language": String,
  "description": String
}
</report_structure>
<report_output_example>
{
  "title": "This is the title of the report",
  "topics": ["topic 1", "topic 2"],
  "language": "EN",
  "description": "This is the description of the report"
}
</report_output_example>
</report>
</instructions>

Benefits

Using Markdown Formatting and Tags makes your instructions more readable and easier to understand for the AI assistant.

Using Hierarchical Structures helps the AI assistant to understand the instructions and to follow them, it's much more effective than having your instructions in a normal plain text.

You can also use them together to get the best of both worlds.