Copy of Learn AI Guide

Hey friend,

Thanks so much for downloading this guide. I genuinely appreciate you being here 🙂

AI is moving fast and I know it can feel overwhelming. But here’s the good news: you don’t need to learn to code, you don’t need a computer science degree, and you definitely don’t need to keep up with every new tool that launches. You just need a clear roadmap and the willingness to start.

That’s exactly what this guide is. It’s a practical, phase-by-phase plan to take you from zero to genuinely productive with AI in about three months. The same framework I use with my own team.

If you haven’t already, I’d really recommend watching the full video that accompanies this guide. It walks through everything in detail with loads of real examples, so you can see how this stuff actually works in practice.

👉 Watch the video here: How to Learn AI

You’ve got this. Seriously. The fact that you’ve downloaded this guide already puts you ahead of most people. Now let’s get into it.

Ali xx

Phase 1: Build Your Foundations (Week 1)

Before you try anything clever, you need the right habits and tools in place. Think of these as non-negotiables.

1. Replace Google with AI

Use an AI chatbot (Claude, ChatGPT, Gemini, Grok, Perplexity… whatever you have access to) for everything you’d normally Google. This is the single biggest mindset shift.

2. Pin It in Your Browser

Keep your AI tool open in a pinned tab at all times. The goal is to build the habit of always having an AI chat window available, no matter what you’re working on.

3. Talk, Don’t Type

Use voice input instead of your keyboard. Most AI tools now have built-in dictation, and you can also use tools like Wispr Flow or the built-in dictation on Mac and Windows. You’ll find you can communicate faster, ramble more freely, and get significantly more value from AI this way.

4. Download the Mobile Apps

Having AI on your phone means you can use it anywhere — while walking, commuting, or just thinking through a problem away from your desk.

5. Auto-Record Your Meetings

Use tools like Grain or Fathom to automatically record and transcribe your Zoom calls, Google Meets, and other online meetings. These transcripts become incredibly useful fuel for AI later on.

⏸️	Pause here. If you haven’t done all five of these things, do them now before moving on. Everything else builds on this foundation.

Phase 2: AI as Your Coach (Week 2)

Now you’ll start using AI in a meaningful way — but you’re not asking it to do your work yet. You’re asking it to help you think better about the work you’re already doing.

How to Use AI as a Thinking Partner

Ask AI to coach you on your specific role. Here are some example prompts:

“I am a [YOUR ROLE] tasked with [YOUR GOAL]. What are the highest-leverage things I should focus on? What mistakes do people in my role commonly make? What questions should I be asking my manager to make sure I’m set up for success?”

“I want you to interview me about what I actually do in my role and help me identify what’s high leverage and what’s probably a waste of time.”

Use Your Meeting Transcripts

Remember those auto-recorded meetings from Phase 1? Now they become gold. Take a transcript and ask:

“Based on this conversation, can you suggest a learning curriculum for me to follow over the next two weeks?”
“Based on this transcript, tease out the key themes and struggles so I can improve my approach.”
“Give me feedback on my communication style and any blind spots you notice.”

⚠️	A word of caution: Think of AI as a very smart colleague who reads a lot of books but has zero real-world context beyond what you give it. Its advice can be incredibly useful for mirroring your thoughts and pushing your thinking, but don’t treat it as gospel. Always apply your own judgement.

By the end of Week 2, you should have the habit of turning to AI whenever you’re stuck — and even when you’re not, just to optimise your performance.

Phase 3: AI as Your Worker (Weeks 3–4)

This is where you start getting AI to do things for you. But there’s a right way and a wrong way.

The 10-80-10 Rule

The biggest mistake people make is asking AI to do 100% of the work. That’s how you end up with generic garbage. Instead, follow the 10-80-10 Rule:

You do the first 10% — Gather context, set direction, provide examples
AI does the middle 80% — Generate the bulk of the output
You do the final 10% — Quality check, taste check, refinement

Example: Generating Content Ideas

❌ The wrong way: “Come up with 50 content ideas for my Instagram.”

✅ The right way: “Here’s a transcript from a recent video. Here are three competitor posts that performed well this month. Here is our content strategy doc with our target audience and brand voice. Based on all this, give me 20 hook ideas. Focus on counterintuitive takes and pattern interrupts.”

The Feedback Loop

Once you get results back from the AI:

Review the output and pick the ones you actually like
Paste your favourites back and say: “These were the five I liked most. Give me 50 more along this vein.”
Repeat until you have a solid set of ideas — all human-vetted

Develop Your Taste

The key that separates people who use AI well from those who don’t is taste: your intuitive feel for what’s good and what isn’t. If the AI’s output makes you cringe, that’s actually a great sign. It means your quality bar is higher than the AI’s default output, and your job is to give it feedback (just like you would a junior team member) until it meets your standard.

Phase 4: AI as a System (Months 2–3)

Up to this point, you’ve been starting from scratch every time you open a new AI chat. Phase 4 is about making AI get better over time through prompt engineering.

Build Your Prompt Library

Think of prompts like recipes. Each time you use one, you refine it:

V1: “Give me 50 Instagram hook ideas from this transcript.”
V2: Add “Make sure each hook uses a pattern interrupt or controversial take. Avoid generic advice.”
V3: Add “Make sure each hook is under 20 words.”
V4: Add “Never use rhetorical questions.”

Each version gets better because you’re encoding your taste and preferences directly into the prompt.

Set Up Shortcuts

Use a tool like TextExpander to create keyboard shortcuts for your best prompts. Type a short code, and the full prompt expands automatically — saving you time and ensuring consistency.

Experiment with Models

Once you have a solid prompt library, test different AI models (ChatGPT, Claude, Gemini) with the same prompts. You’ll find that certain models work better for certain tasks. At this stage, a paid subscription to one or two AI tools is usually well worth it.

Explore Specialised AI Tools

Depending on your work, you might benefit from AI tools built for specific tasks — slide deck generators (Gamma, Beautiful.ai, Figma Slides), image generators, writing assistants, and more. Don’t get overwhelmed by choice — just find the tools that serve your specific use cases.

Phase 5: AI as Infrastructure (Month 4+)

This is where you stop being in the loop entirely for repetitive tasks and let AI run in the background.

Level 1: Built-In AI Features

Use AI automation that’s already baked into tools you’re using. Many editing, transcription, and project management tools now have AI features that automate steps you were doing manually.

Level 2: Simple Automation Tools

Use tools like Zapier or Make to connect apps together. For example:

Every time there’s a new Zoom recording → auto-transcribe → run through a ChatGPT prompt → send the output as a Slack message

Level 3: Advanced Automation

Graduate to more powerful tools like n8n for more granular, sophisticated workflows. These require a bit more technical knowledge but unlock far more complex automations.

Level 4: Build Your Own Tools

At the most advanced level, you can build custom internal AI apps for your team or business. But honestly, most people won’t need to go this far — Levels 1–2 cover the majority of use cases.

🧠	The real discipline at this stage isn’t figuring out what to automate — it’s deciding what’s actually worth automating vs. what you should keep doing manually vs. what you can just stop doing entirely.

Quick-Start Checklist

The key takeaway: don’t try to skip ahead. Each phase builds on the last. Get the foundations right, develop your taste, build your systems, and you’ll be genuinely fluent with AI within a few months.

Software Engineering Life

Explorer