The Method
First, pick your AI.
You do not need the “best” model. You need one good assistant you will actually use. The concepts in this course transfer between products, so choose the one that fits your world and start talking to it today.
A balanced first AI for learning, everyday questions, files, voice, and general building.
A strong default if you want a broad, general-purpose starting point.Open ChatGPT ↗
Long explanations, writing, reasoning through documents, and learning while you build.
A strong fit if you like thoughtful, longer-form collaboration and explanation.Open Claude ↗
People who already live in Google products and want their AI experience close to that ecosystem.
A natural starting point when Gmail, Drive, Docs, and other Google tools are already your home base.Open Gemini ↗
People especially interested in current internet conversation, trends, and the X ecosystem.
A natural fit when live web context and the X ecosystem matter most to you.Open Grok ↗If your employer already gives you an approved AI assistant, start there. Do not paste confidential, customer, company, credential, or private data into a personal AI account just to follow this course.
Which one should I start with?
This is a fit quiz, not a benchmark. Pick the answer that sounds most like you.
Do not explore every button. Start a conversation.
Paste this, or say it out loud:
I am brand new to using AI as a serious tool. Ask me one question at a time about what I do, what I already know, what I want to learn or build, and how I like to learn. When you understand me, build a practical learning path and give me only the first three things to do.
The fastest way to learn AI is not to memorize tools. It is to build a loop where AI learns your context while you learn the system: ask, do, save, automate, and ask again.
"My honest take is: literally just ask your AI. Use the voice feature to explain everything that you do at work. Talk about your role. Talk about the level of things you already know. Then talk about where you want to get to. And have it build the plan."
Before any tool, any line of code, any deployment — this is the method. Everything else in this series is just making you fluent enough to do this well.
Why "just ask" actually works now
For most of history, learning a technical skill meant: find a course, follow someone else's path, hope it overlaps with your job, get stuck, and have nobody to ask at 9pm.
That's over. You now have, available 24/7, a patient expert that:
- Knows your tools (SQL, Tableau, Teams, VS Code, Salesforce, the whole stack)
- Will explain the same thing five different ways until it clicks
- Never judges the "dumb" question
- Can look at your actual files and your actual data and teach from those
- Can remember what you've covered and build on it tomorrow
The bottleneck is no longer access to knowledge. It's knowing how to have the conversation. So let's learn that.
The conversation, step by step
Open the AI assistant you chose above. If you can, turn on voice and just talk. Talking lowers the bar — you'll explain more, and more honestly, than you'd ever type.
Step 1 — Tell it who you are and what you do
"I'm a [analyst / engineer / PM] at an energy company. My job is mostly [pulling job-profitability data, building Tableau dashboards, answering ad-hoc questions from operations]. I spend my day in Outlook, Teams, Tableau, Excel, and SQL Server Management Studio."
The more specific, the better. Your AI tailors everything to your context.
Step 2 — Tell it your real skill level
Don't undersell or oversell. Be honest:
"I can write a SELECT query and join two tables. I've never used Git. I've opened a terminal maybe twice. I don't know what an API is, but I've heard the word a hundred times."
This is the most valuable sentence you'll say. It's the difference between your AI talking over your head or wasting your time on stuff you know.
Step 3 — Tell it where you want to go
"In six months I want to be able to build a little internal tool that pulls our frac job data and shows it as a dashboard, without waiting on IT. Eventually I'd love to build a bot that answers the team's repeat questions."
Step 4 — Ask for the plan
"Based on where I am and where I want to go, build me a custom lesson plan. Order it so each step builds on the last. Keep it hands-on — I learn by doing, not by reading."
Now you have a curriculum that no course could ever match, because it starts exactly where you are.
Step 5 — Learn by doing, and narrate
As you work, keep talking:
"Okay I ran that and got an error that says ORA-12541. What does that mean? Explain it like I've never seen it.""Wait, why did we use a virtual environment? What problem does that solve?"
Every error is a lesson. Every "why did we do that?" is a lesson. Your AI is happiest answering exactly these.
The three things that turn "asking" into a system
A one-off chat is helpful. But the magic compounds when you do three things — and these get their own modules:
- Save what you learn → Memory (Module 04).
Tell your AI: "Remember that I'm on SQL Server SQLSERVER01\REPORTING, the main database is OperationsData, and I prefer explanations with a real example." Now every future session already knows your world. You're not starting from zero each morning.
- Turn repeat tasks into Skills (Module 05).
The report you build every Monday? The five-step thing you always forget the order of? Have your AI package it into a skill — a saved, named workflow you can trigger with one command. You teach it once; it remembers forever.
- Have it curate and adjust the plan as you grow.
Come back weekly: "Here's what I got comfortable with this week. Update my lesson plan and push me a little further." The curriculum is alive.
This is the whole game: ask → do → save → automate → ask again. Memory and skills are what separate "I used AI once" from "AI is now part of how I work."
What this is not
- Not "let the AI do it so I don't have to learn." You'll learn faster this way, not less. The point is to understand what you're building.
- Not a replacement for understanding the fundamentals. The next modules give you the mental models (what's an LLM? what's an API? what's a token?) so that when you ask your AI something, you can follow the answer and catch it when it's wrong.
- Not a license to point experiments at production systems. Learn on safe data. (Especially: leave Oracle alone until you really know what you're doing.)
🎯 Try with your AI (do this before moving on)
First, give your AI a map of the course so it can understand what you are learning and follow along as your study partner.
What is a Markdown file?
Markdown is a plain-text file that uses simple symbols for headings, lists, links, and emphasis. Its file extension is .md. AI assistants work especially well with Markdown because the structure is clear, lightweight, and easy to scan without hidden formatting.
Download the AI Coffee Shop course companion (.md) ↓
Introduce the course to your AI
- Download the course companion above.
- Open the AI assistant you chose and attach the
.mdfile. - Paste the prompt below. Turn on voice if you can, then continue the conversation honestly.
I am working through the AI Coffee Shop course and I am just getting started. I attached the course content so you can follow along with me. Please act as my learning partner throughout the course. Start by asking me one question at a time about my work, what I already know, what I want to learn or build, and how I prefer to learn. Use my answers and the course guide to help me connect each module to my real goals. Do not overwhelm me—when we make a plan, give me only the first three actions to take.
Spend five honest minutes answering the questions. When the AI understands you, ask it to build a practical learning path and save that plan somewhere you can return to it.
AI is most useful as a learning loop, not a one-shot answer machine.