A Playbook Is Just a Manual A playbook is just a manual for how you do your work, written in a way that can be followed consistently by either you or an AI system. It breaks a task into clear, structured steps so nothing is left to interpretation: Step one Step two Step three Step four But the key part most people miss is this. A playbook is not just instructions, it also needs trigger conditions. Without triggers, the system has no idea when to activate, which means it either runs at the wrong time or not at all. So alongside your steps, you define a simple activation rule, for example: Run flyer maker Now the system knows exactly when to execute that playbook, and when to ignore it. This is already being used in systems like Claude Skills, where specific behaviours are triggered when certain conditions are met in conversation. If the user talks about newsletters, the system recognises it and activates the relevant skill automatically. The same logic applies across modern AI systems. Custom GPTs, projects, and Gems all rely on the same idea. You are no longer just building prompts, you are building structured behaviours. For example, you could create a single Gem that contains multiple playbooks inside it: A newsletter playbook triggered by “run newsletter maker” A flyer playbook triggered by “run flyer maker” A copywriting playbook triggered by “run copywriter” Each one sits inside the same system, but only activates when its specific trigger condition is met. This is where it becomes powerful. You are not juggling separate tools, you are building one structured system that routes tasks automatically based on intent. Even on something as simple as a free Google account, you can store up to 10 documents. That is 10 separate playbooks, each one capable of handling a full system of work. If you use Google Drive alongside something like a Gem, it becomes even more flexible. You can connect those documents directly, meaning any update you make inside Drive immediately improves the behaviour of your system.

🧠 Playbooking Method Training Guide How to turn repetitive work into AI-powered systems 1. Core Idea Lisa: A playbook is just a manual for AI. Instead of prompting AI from scratch every time, you pre-build a structured instruction system that defines how a task should always be done. Think: Human SOP → AI Playbook One-off prompting → Repeatable system Guessing → Defined process 2. The 4 Building Blocks of Every Playbook Every playbook is built from the same structure. 1. Trigger What activates the playbook? Examples: Every Monday at 9am After a client call When an email arrives When a task is added Expected result: AI knows when to run without re-explanation. Common errors: Trigger is too vague (“when needed”) No clear activation condition 2. Inputs What does AI need every time? Examples: Meeting transcript Brain dump Email content Task description Expected result: AI knows exactly what information it will receive. Common errors: Missing inputs Assuming AI “knows context” Not specifying format (messy input = messy output) 3. Steps What is the process the AI follows? Example (newsletter): Pick topic Create outline Write draft Edit in brand voice Generate subject line Expected result: Consistent output quality every time. Common errors: Steps too broad (“write newsletter”) Steps not ordered logically Mixing strategy and execution in same step 4. Output What is the final result? Examples: Drafted newsletter Client summary Email response set Marketing campaign plan Expected result: Clear finish line for AI. Common errors: No defined output format Multiple outputs with no structure 3. Making the Playbook Work Properly Lisa: This is where most people fail. The structure alone is not enough. You must add depth. Add: Clear instructions (no contradictions) Context (who, why, situation) Examples (what good looks like) Success criteria (checklist of quality) Success Criteria Example For a newsletter: Clear and casual tone One main idea only