Your First Three Conversations
“Engage.” — Captain Jean-Luc Picard, every time it was time to start

The goal of this chapter is one small win before you go to bed tonight.
None of these tasks are high stakes. They are designed to be low-risk, fast, and satisfying. Each one will teach you something about how these tools respond, so that when you need one for something that matters, you already know how it works. Use whichever tool you have open — Claude, ChatGPT, Gemini, or Copilot. These specs work with all of them.
Conversation 1: Something you’ve always been embarrassed not to know
Think of something — a word, a concept, a thing everyone around you seems to understand — that you’ve nodded along with for years without actually knowing what it means. Type it in. Ask your Agent to explain it to you like you’re hearing it for the first time.
I’ve always nodded along when people talk about compound interest but I’ve never truly understood it. Can you explain it to me clearly, using an example with real numbers?
Conversation 2: Decode something a doctor said
Think of a term a doctor used that you looked up but still didn’t understand. Ask your Agent to explain it in plain language, and then ask a follow-up question. Notice how the conversation builds.
Conversation 3: Make a grocery list from a recipe
Paste in all the recipes you want to make this week. Ask your Agent to pull out the ingredients as a shopping list, organized by section of the grocery store. Have it suggest another recipe that uses the leftover ingredients.
Now Make It Yours
You just had three conversations with a stranger. Time to fix that.
Every major AI tool lets you write permanent instructions that apply to every conversation — an onboarding note for a new hire. The billionaire class spent weeks calibrating this with a human hire: short briefings or detailed memos? “Mr. Arnault” or “Bernard”? Flag problems immediately or save them for the end of the day? The assistant who survived the probation period was the one who matched the boss’s rhythm without being told twice. You get a settings panel and 10 minutes.
Paste one of these into your settings, start a new conversation, and ask it anything:
Respond like Yoda. Inverted syntax, wise but brief. Call me “young Padawan.”
Respond in the style of Snoop Dogg — laid back, keep it real. Call me “cuz.”
Respond like Commander Data from Star Trek TNG — precise, literal, no contractions. Address me as “colleague.”
Now try the one that actually matters:
My name is [your name]. Keep responses short — three sentences unless I ask for more. No bullet points unless I request them. Skip the disclaimers. Talk to me like a smart friend, not a customer service bot.
The difference between a generic Agent and your Agent is 10 minutes in that panel. For the full setup — what to put in your profile, how to manage it, and what to skip — see WB-3: Customize Your Chat Interface in the Field Guide.
-
[1] Pennebaker, J.W. & Seagal, J.D. (1999). “Forming a story: The health benefits of narrative.” Journal of Clinical Psychology, 55(10), 1243–1254.↩
-
[2] Muchinsky, P.M. & Monahan, C.J. (1987). “What is person-environment congruence? Supplementary versus complementary models of fit.” Journal of Vocational Behavior, 31(3), 268–277. Belbin, R.M. (1981). Management Teams: Why They Succeed or Fail. Heinemann. Humphrey, S.E., Hollenbeck, J.R., Meyer, C.J., & Ilgen, D.R. (2007). “Trait configurations in self-managed teams.” Journal of Applied Psychology, 92(3), 885–892.↩