How to read this manual

This is how you can have your agent do all of your digital work for you, so you interface with your tools far less often. You just talk to the agent and the agent does the work. After reading this you should be able to: Tell your agent what to do without overthinking the prompt. Make rules that actually stick instead of getting forgotten. Build skills so you only teach it once. Connect tools (email, calendar, CRM, browser, files) without breaking anything. Fix it yourself when something goes sideways. Spot when you've gone from "I don't get this" to "this is doing my job for me." Read straight through if you are new. Skim the table of contents and jump if you have a specific problem. The troubleshooting and security chapters are the ones people come back to most.

Table of contents

1. What an AI agent actually is

1. Your first day with the agent
2. How to talk to it (prompting basics)
3. The magic phrases
4. Memory and always-loaded context
5. Skills — teach it once, run forever
6. Daily habits of power users
7. Braindumps — stop writing notes, start talking
8. Training it on your voice
9. When to use sub-agents
10. Making the agent proactive
11. Connecting tools (email, calendar, CRM, browser)
12. Communication channels
13. What the agent remembers and how
14. Troubleshooting — fix anything in five minutes
15. Security and privacy
16. Hosting choices and keeping costs down
17. The click moments — when you become dangerous

19. Common mistakes (and how to skip them)

Glossary at the end. How to use this book without reading it cover-to-cover, and how to hand it to your agent.