---
id: "framework-parenting-ai"
type: "framework"
source_timestamps: ["01:25:50"]
tags: ["prompt-engineering", "best-practices"]
related: ["concept-parenting-ai", "quote-parenting-ai", "action-build-tone-mirror"]
steps: ["Provide Context: Tell the AI exactly who you are, what your business is, and what you care about.", "Provide Examples: Show the AI what 'good' looks like by pasting in examples of your previous work, writing style, or desired formatting.", "Provide Iterative Feedback: When the AI generates an output, do not accept the first draft if it's flawed. Tell it specifically what is wrong (e.g., 'That's too formal, make it more conversational') and have it try again."]
---
# Parenting AI (Prompting Framework)

[[entity-lior-weinstein|Lior Weinstein]]'s framework for interacting with LLMs. Instead of trying to write complex code, treat the AI like a new employee or a child that needs context and guidance — see [[concept-parenting-ai]].

## Steps

1. **Provide Context** — tell the AI exactly who you are, what your business is, and what you care about.
2. **Provide Examples** — show the AI what *"good"* looks like by pasting in examples of your previous work, writing style, or desired formatting.
3. **Provide Iterative Feedback** — when the AI generates an output, do **not** accept the first draft if it's flawed. Tell it specifically what is wrong (e.g., "That's too formal, make it more conversational") and have it try again.

## Slogan

> [[quote-parenting-ai|"Don't program it. Parent it."]]

## Worked Example

The [[action-build-tone-mirror|Tone Mirror prompt]] is a complete worked instance of this framework — feeding the AI your style, your values, and inviting iterative critique.