---
id: "framework-playbook-outline"
type: "framework"
source_timestamps: ["01:18:00"]
tags: ["automation", "process-design"]
related: ["concept-playbooking-method", "concept-operational-debt", "action-create-playbook", "entity-claude", "entity-zapier"]
steps: ["Trigger: Define the initiating event.", "Inputs: Identify the required variable data.", "Steps: Break the process into sequential tasks.", "Outputs: Define the expected final deliverable."]
speakers: ["Rachel Woods"]
sources: ["day3"]
sourceVaultSlug: "ai-advantage-summit-2026-2026Apr26"
originDay: 3
---
# The Playbook Outline

## Purpose

A four-step framework for translating a human workflow into a set of instructions an AI agent can execute autonomously. This is the operational core of the [[concept-playbooking-method]] and the practical instrument for paying down [[concept-operational-debt]].

## The four steps

### 1. Trigger
Define **exactly what event** kicks off the playbook.
*Examples:* "Every Monday at 9 AM," "When a new lead form is submitted," "When a meeting transcript is saved."

### 2. Inputs
Identify the **variable data** the AI needs each time it runs.
*Examples:* a brain dump of ideas, a meeting transcript, lead contact info, the week's sales numbers.

### 3. Steps
Break down the actual process into **sequential, specific tasks.**
*Example sequence (newsletter playbook):*
1. Pick a topic.
2. Outline.
3. Draft.
4. Edit for brand voice.

### 4. Outputs
Define the **exact end result** expected from the AI.
*Example:* "A drafted newsletter ready for human review, in markdown, ≤ 500 words, in our brand voice."

## Where the playbook lives

Load the document into a tool such as [[entity-claude]] (preferred for large context windows and brand-voice fidelity), a custom GPT, or [[entity-zapier]] (for fully automated background execution).

## The action you should take

Do this for one task, this week — see [[action-create-playbook]].

## Enrichment context

Conceptually similar to LangChain agent toolkits, Auto-GPT playbooks, and Hugging Face's *Building AI Agents* documentation. The Trigger / Inputs / Steps / Outputs schema is unusually clean for a non-technical audience and serves as a gentle on-ramp to agentic patterns.


## Related across days
- [[framework-parenting-ai]]
- [[concept-agentic-ai]]
- [[concept-playbooking-method]]
- [[concept-active-level-3]]
- [[arc-agentic-progression-from-oracle-to-playbook]]