# AI 101 · Lesson 2 — Your AI roadmap: what to learn, in what order

**The CV Method · AI 101 · Lesson 2 · Working with AI · Aug 2026**
*by Corina · @the_cv_method · hello.thecvmethod.com · current as of Aug 2026 — the AI space moves fast, so double-check tool names*

The overwhelm isn't the tools — it's not knowing what comes next. Here's the whole beginner path on one page, so you go from "AI is a lot" to "I use it every day" **without drowning**. Six stops. No skipping ahead.

---

## The map (at a glance)

1. **Just talk to it** — use AI on one real task, today
2. **Learn to prompt** — it's structure, not magic words
3. **Give it your context** — the docs that make it reliable
4. **Automate one boring task** — your first tiny "agent"
5. **Connect your tools** — one at a time, when a task needs it
6. **Put it on a schedule** — let the reliable stuff run itself

## Rule zero — learn it on ONE real task at a time. Depth beats breadth.

You don't learn AI by watching forty tutorials — you learn it by pointing it at one real thing this week. Master a stop, then take the next one. That's the whole method.

## Stop 1 · Just talk to it

Before anything fancy, just **use it**. Pick one real task you'd do anyway — draft an email, summarize a long doc, rewrite a rambly paragraph — and hand it to the AI. The goal isn't a perfect result; it's getting a feel for what it's good at (and where it needs you).

*You're ready for Stop 2 when: talking to it feels normal, and you're mildly annoyed that you keep re-explaining yourself.*

## Stop 2 · Learn to prompt

Good prompting isn't secret phrases — it's just being clear. Tell it four things: the **role** ("you're my assistant"), the **goal** ("draft a reply that does X"), the **constraints** ("short, friendly, no jargon"), and **an example** of what good looks like. That's 90% of it.

*You're ready for Stop 3 when: your first draft is usually close, and you find yourself pasting the same background info again and again.*

## Stop 3 · Give it your context

This is where Lesson 1 pays off: put your background in a **CLAUDE.md** and the relevant files in a folder, and point the AI at them. Now it knows who you are and how you work **every time** — no more re-typing your life story. Good context is the difference between "generic" and "sounds like me."

*You're ready for Stop 4 when: it reliably nails your voice and rules, and there's a boring task you wish would just do itself. (New here? Do Lesson 1 first — "Set up your AI stack.")*

## Stop 4 · Automate one boring task

Now take **one** small, repeatable task and set the AI up to do just that — "draft my weekly update," "turn this into a to-do list." That's all an "agent" is: one job, the right context, and rules (it drafts, you approve). Small and boring is the goal.

*You're ready for Stop 5 when: that one task is reliably handled, and it'd be even better if the AI could reach one of your tools. (Want the full how-to? That's "Set up your first AI agent.")*

## Stop 5 · Connect your tools

When a task genuinely needs it, connect **one** tool — your email, calendar, notes, or docs — so the AI can read or draft in the real thing. Add the next connection only when the next task asks. Every connection is power **and** a new thing to keep an eye on, so go slow and keep the "it drafts, you approve" rule.

*You're ready for Stop 6 when: a task works end-to-end by hand and you're doing it on a predictable schedule anyway.*

## Stop 6 · Put it on a schedule

Once a task is boringly reliable by hand, let it run on a schedule — a draft waiting for you each morning. Build in a **fail-safe** (if it can't do the job, it pings you instead of guessing), and it still **never sends on its own**. This is the payoff: your time back, judgment kept.

*You've arrived. From here it's just more of the same — one new task, one new stop, on repeat.*

## How to know you're ready to move on

Simple rule: **move on when the current stop feels boring and easy** — not before. If a stop still feels hard, stay there and do it on another real task. Rushing ahead is the fastest way back to overwhelmed. Slow is smooth, and smooth is how this sticks.

## Keep it honest & safe

- One task at a time.
- It proposes, you approve.
- Move on only when it's easy.
- Keep private info out.
- No invented numbers.
- Date it — AI moves fast.

## Steal this prompt

**Find your next stop — paste and fill in**

```
I'm a beginner learning to use AI. Here's what I can already do: [a sentence or two].
Based on that, what's the ONE next thing I should learn — and a small, real task from
my week I could practice it on? Keep it plain English, one step, no overwhelm.
```

## The roadmap checklist

1. **Stop 1:** use AI on one real task today.
2. **Stop 2:** prompt with role · goal · constraints · example.
3. **Stop 3:** give it a CLAUDE.md + the right files.
4. **Stop 4:** automate one boring task (it drafts, you approve).
5. **Stop 5:** connect one tool, when a task needs it.
6. **Stop 6:** schedule the reliable stuff, with a fail-safe.
7. Move on only when a stop feels **boring & easy**.
8. One real task at a time — **no skipping ahead**.

---

**Want the whole roadmap walked with you — or built for your business?** I help beginners find their next step, and I build the whole AI foundation for businesses that would rather skip to "it just works." Start at **hello.thecvmethod.com**.

*The CV Method · AI 101 · Lesson 2 — Your AI roadmap*

---

*Part of the **AI 101** series: Lesson 1 — Set up your AI stack · Lesson 2 — Your AI roadmap · Lesson 3 — How to prompt · Lesson 4 — Give it your context. More lessons at hello.thecvmethod.com/playbooks.*