The Non-Technical Guide to AI

You don't need to
be technical to start
using AI today.

AI is the most powerful tool available to you right now. The only thing standing between you and it is starting. Here's how.

Show me how → See the tools
1-2 hrs
Per week is all you need to start seeing a massive difference
~$20
Per month for a Pro subscription — the single best ROI you can make
Step by Step

How to eat the elephant 🐘
one small bite at a time.

You don't have to be good at AI. You just have to start. Here's a simple path to go from zero to genuinely useful in a matter of weeks.

01

Start free, upgrade when you're ready

Start with a free ChatGPT or Claude account. If you can upgrade to Pro (~$20/month), even better — free tiers are often slightly behind, but don't let it be a barrier to entry. ChatGPT is great for creative comms and stakeholder writing. Claude is your go-to for analysis, research, and thinking through complex problems. Pick one, commit for a month.

Check out easy prompts to try →

Week 1
02

Put it on your calendar like a meeting

Create a calendar rock — 1–2 hours every Friday morning, or two 1-hour blocks per week — whatever works for you. Treat it like a ritual, not a task. The goal isn't to be perfect; it's to make playing with AI a habit you actually keep. Consider creating a shared channel (like #ai-discuss) with friends or teammates to build accountability.

Week 1–2
03

Find a pain-point and do something about it

Create a list of manual things you do in your job.

Don't just read about AI — use it for a real task you have this week. Draft that email you've been putting off. Ask it to summarize a report. Have it write a plan for a project. Automate a workflow (ask Claude how and it will guide you!) The fastest way to understand what AI can do is to see it solve a problem you actually have.

Week 2
04

Build something you can show someone

Use Gamma to turn a document into a beautiful presentation. Use Lovable/Replit to build a simple webpage. Create a deep research report in Claude on a topic you care about. Generate an image from text in GPT or Nanobanana. The moment you build something you'd actually share — a birthday page, a beautified headshot, a pitch deck, a research report — is the moment AI clicks for you.

Week 3–4
05

Watch what others are building

The fastest way to level up is to see what's possible. Spend time on YouTube and X (Twitter) watching people build in public — practical walkthroughs, new tools, lessons learned. You'll constantly discover new ideas to try on yourself. Channels like How I AI, the a16z Newsletter and thought leaders/builders like Andrej Karpathy and Olivia Moore are great starting points.

Week 3–4
06

Connect your apps via MCPs

MCP (Model Context Protocol) lets your AI connect directly to the tools you already use — think email, Slack, Google Drive, Notion, calendars. Instead of copy-pasting between apps, your AI can read, write, and act across all of them. Ask Claude to set it up and it will walk you through it step by step.

Note 🤓: if you're doing this at work, confirm you're on an enterprise account with the right security protocols.

Month 2+
07

Practice the art of prompting

A prompt is the instruction you give to an AI — and the quality of what you get back is almost entirely determined by how well you ask. A good prompt gives context, assigns a role, specifies the format you want, and is specific about the outcome. If you can share an example of the output, even better.

Prompt Cowboy is a fun way to turn lazy prompts into better ones. 🤠

Month 2+
08

Build AI into your workflow

Create a custom GPT trained on the topics you query most. Set up repeatable prompts for tasks you do regularly. Have questions you always research before a call? Have a format in which you want communications written? GPTs are great for this (or Claude Projects, Google Gems).

Month 2+
09

Build an agent to automate a process

An agent is an AI that can take a sequence of actions on your behalf — without you directing every step. Pick a repetitive workflow (weekly reports, email triage, data entry) and automate it. Tools like Zapier, n8n, Gumloop, and DUST make this accessible without code. If you have a paid Claude account try CoWork. You can set all of these up using natural language (no coding required!)

Month 3+

The Toolkit

Apps worth trying
right now.

You don't need all of these. Pick one from each category, try it for a week, and see what sticks.

💬

ChatGPT

The best all-rounder. Great for drafting comms, creative writing, stakeholder updates, and brainstorming. Go Pro.

Chat & Writing
🧠

Claude

Exceptional at analysis, long documents, nuanced reasoning, and building things with code. Your research co-pilot.

Analysis & Research
🎨

Gamma

Upload a doc or type a prompt and get a beautiful presentation or webpage in seconds. The easiest entry point for building.

Presentations
❤️

Lovable

Build real web apps and websites from templates without writing a single line of code. Genuinely magical.

No-code Apps
🎵

ElevenLabs

Create AI voiceovers, podcasts, and soundtracks. Also great for building your own voice avatar.

Audio & Voice

n8n

Automate workflows and connect your apps using natural language — no code required. Describe what you want to happen and let it build the process for you.

Automation
🖼️

Ideogram or Reve

Generate and remix images from text prompts. Great starting point — many apps provide example prompts to inspire you.

Images
🗣️

Sesame (Miles)

Have a real voice conversation with an AI. The first time you try this, it will blow your mind.

Conversational AI

"The best prompt is like a wish to a genie — the outcome is only as good as the ask. Be specific, give context, and say what you actually want."
The #1 thing that makes AI more useful, immediately
Example Prompts

Copy, paste, and
make it your own.

These are some easy prompts to get you started seeing AI at its best — replace the bracketed parts with your own context and watch what happens. Some ideas of what to build in the people space here →

Process & Ops

"Here is our current process for [X]: [paste steps]. Identify the inefficiencies, suggest where AI could automate or assist, and give me a phased implementation plan."

Decision Making

"I need to make a decision about [X]. Here are the options and what I know: [paste]. Build me a decision matrix, weight the criteria, and give me a recommendation with your reasoning."

Strategy

"Play the role of a smart, skeptical leader in [X discipline] and pressure-test this business idea: [paste]. Don't be kind — find the holes."

Performance

"Act as an operations consultant. Here are our team's OKRs and last quarter's results: [paste]. Give me an honest assessment of where we're off track and why."

Research & Insight

"Analyse this survey data and identify the top 3 themes, any red flags, and recommended actions. Then suggest three things the leadership team should prioritise in the next 90 days based on what you find: [paste data]"

People & HR

"I'm about to have a difficult conversation with an employee about performance. Here's the context: [paste]. Help me plan what to say, anticipate their likely responses, and suggest how to keep it constructive."


Key Terms

The words you'll hear
and what they mean.

You don't need to memorise these. Just know enough to follow the conversation.

Prompt
The input you give to an AI model. Think of it as your wish to a genie — the more specific and clear, the better the output.
LLM
Large Language Model. The AI brain behind tools like ChatGPT and Claude. Trained on vast amounts of text to understand and generate language.
RAG
Retrieval Augmented Generation. A way of giving AI access to your own documents or data, so it can answer questions using your specific context.
Fine-tuning
Additional training on specialised data to make a general AI model an expert in a specific task — like teaching a generalist to specialise.
Vibe Coding
Building software by describing what you want in plain language, without writing code yourself. Coined by Andrej Karpathy.
MCP
Model Context Protocol. A way of connecting AI to the tools and data you already use — like email, Slack, or Google Drive. Think of it as an API for the AI world.
Chain of Thought
A prompting technique where you ask AI to reason step-by-step. Seeing how the model thinks through a problem dramatically improves accuracy.
GPU / TPU
Specialised hardware that powers AI. GPUs (NVIDIA) handle parallel processing for AI training. TPUs are Google's custom-built ML chips.
Agent
An AI that takes multi-step actions autonomously — browsing the web, writing code, sending emails — without you directing each step.

Your Challenge

Build something by
end of next week.

The best way to learn is to commit to making something real. One small challenge per day — by Friday you'll have a completely different relationship with AI.

The 5-Day AI Sprint 🚀

One small challenge per day. By Friday, you'll have five real things built and a completely different relationship with AI tools.

Monday Use Deep Research (GPT or Claude) to create a report on a topic you care about
Tuesday Remix a photo or create an image from scratch using Ideogram or Reve
Wednesday Turn a document you already have into a Gamma presentation
Thursday Have a voice conversation with Sesame — just see what happens
Friday Build a simple webpage on Lovable for something real in your life