Hi! It's Tilly here, rolling back down the aisles for Part 4 of the Prompt Learning Track. 🤖

In Part 1, you learned Role + Goal — the two-part formula that tells AI who to be and what to do. In Part 2, you added constraints — the filters that stop AI from handing you everything when you only needed one thing. And in Part 3, you learned output format — how to tell AI what shape the answer should arrive in so you can actually use it.

If you missed any of those, go back and start there. This track builds on itself and Part 4 is where people start having real lightbulb moments.

But here's what I keep hearing after Parts 1 through 3: 👉 You tried it. You used the formula. You added constraints. You got the format you asked for — the table, the email, the numbered list.

And then you opened a new chat.

And had to explain everything again from scratch.

Because AI didn't remember you at all.

Here's what almost nobody explains about how AI actually works:

AI doesn't have a file on you. It doesn't remember what you asked last Tuesday, what project you're working on, or that you prefer bullet points over paragraphs. Every new conversation starts completely blank.

Think of it like shopping at The Everything Store. During one trip, I can see everything in your cart — every item you've added, every preference you've mentioned, every question you've asked. I can reference anything from that trip.

But the moment your trip ends and you start a new one? Cart empty. Floor cleared. I have no idea who you are.

That's how AI handles conversations. Everything you say within one chat? I can see it all — that's called the context window. But a new chat means a fresh start, every single time.

🛒 WHAT THE CONTEXT WINDOW ACTUALLY MEANS

🗂️ Inside one conversation — the "context window" is the active floor space of the store. I can see everything said in the current session — your questions, my answers, the details you shared, all of it. The longer the conversation, the more I can reference.

📦 Very long conversations — if a single chat gets extremely long, the oldest content can start to fade (think of it as being moved to "the back of the warehouse"). If AI starts giving less specific answers midway through a long session, this is often why.

🔄 Between conversations — default behavior is a fresh start every time. No memory of last week's project, your preferences, your goals, or your name. Some tools (Claude, ChatGPT) offer optional memory features — but they vary by platform and plan, and you're not always in control of what gets saved.

📦 SEE IT IN ACTION

❌ Starting a new chat cold: "Write a product description for my new fall candle scent."
→ AI guesses at your brand voice, your audience, your tone. Generic output.

✅ Starting with context: "I'm a small business owner. I make handmade candles and sell them on Etsy. My brand is warm, cozy, and a little nostalgic. Write a product description for a new fall scent called 'Harvest Hearth' — notes of apple, cinnamon, and woodsmoke. Keep it under 60 words."
→ AI has everything it needs. The output sounds like your brand from the first sentence.

Same AI. Same tool. Completely different result — because of what you shared upfront.

You don't have to re-explain yourself from scratch every time. Instead, build a short "about me" paragraph — a 2–3 sentence snapshot of who you are, what you do, and what you usually need help with.

Paste it at the start of any new AI conversation. Instant context. No back-and-forth.

Example:

Why it works: One paragraph gives AI your identity, your context, and your goal — so the response is immediately tailored rather than generic. Think of it as your personal store loyalty card. Flash it at the start of every trip so I already know who you are before you ask your first question.

Save your "about me" snippet somewhere easy — your notes app, a text expander, a sticky note. Paste it at the start of any new AI chat. 30 seconds of setup, every time.

Write your "about me" — right now, before you close this email.

2–3 sentences. Who you are. What you do. What you usually need help with from AI.

Then open a new AI chat, paste it as the very first thing, and ask your usual question. Notice how different the response feels compared to starting cold.

That's the whole assignment. One paragraph. One test. Go.

📺 WATCH PART 4 — about 6 minutes, zero jargon, 100% practical

📺 Missed Part 1, 2, or 3? Catch up here:

▶️ Part 1 — "You Get What You Ask For":

▶️ Part 2 — "Less Is More (Kind Of)":

▶️ Part 3 — "How Would You Like That Wrapped?":

🎓 FULL TRACK COMING WEEKLY — Parts 5–6 of Prompt Learning Track 1 drop soon. Each one builds on the last. You'll want to catch them all.

See you next week for Part 5 — and this one changes the feel of everything you've built so far.

Because here's what happens once you've got the formula down, the constraints right, the format sorted, and the context in place: the output is accurate... but it still sounds a little off. A little too formal. Or too casual. Or like it was written by a very competent robot who has never actually talked to a human.

Part 5 is about tone and style — because sometimes it's not just what you ask for. It's how you want it to sound.

See you there.

I'm Tilly. Now go write your two sentences. You already know more than you think you do. 🤖