Wonder Conductor cohort starts 5th May 2026· 11 days left

Secure your spot →
Sarah Pirie-Nally

AI for Beginners · By Sarah Pirie-Nally

The AI Stack
as a Kitchen.

Strip the jargon. Map it to something physical. And it clicks fast. Here's the whole AI stack — explained as the kitchen it actually is.

🧠

The Chef

AI

The head chef who knows thousands of recipes.

+

The AI is your head chef. It knows thousands of recipes, can improvise meals on the spot, and understands what you mean when you say "something fresh, quick, and healthy."

It's brilliant at pattern recognition, language, and synthesis — but it doesn't remember everything perfectly. And sometimes it'll confidently cook something… slightly wrong. 😅

❌ Without AI

You're doing everything manually — every recipe, every decision, every word.

✅ With AI

You have a world-class chef on call, 24/7, ready to cook anything you describe.

💡

The honest bit: Most people are talking to the chef but never letting them into the kitchen. That's why results feel limited.

📚

The Recipe Books + Pantry

RAG

Retrieval Augmented Generation — the chef checking the actual recipe.

+

RAG stands for Retrieval Augmented Generation. Fancy name, simple idea: instead of guessing from memory, the chef opens the actual cookbook and checks what's in the pantry before cooking.

This means the meal is accurate, based on real ingredients, and not made up on the fly.

❌ Without RAG

"I think this is how carbonara works…" (Confident. Possibly wrong.)

✅ With RAG

"Let me check the actual recipe and ingredients first." (Accurate. Grounded in reality.)

💡

The honest bit: RAG is why AI can answer questions about your specific documents, your business, your data — not just general knowledge.

🔌

The Kitchen Appliance System

MCP

Model Context Protocol — the system that connects the chef to the tools.

+

MCP (Model Context Protocol) is the system that lets the chef use all the kitchen tools seamlessly. It connects the chef to the oven, blender, fridge, and dishwasher.

Without MCP, the AI can only talk about cooking. With MCP, it can actually use the kitchen — booking your calendar, reading your emails, searching the web, running your CRM.

❌ Without MCP

The chef stands in the middle of the kitchen describing what they'd cook. Nothing actually happens.

✅ With MCP

The chef picks up every tool, uses every appliance, and produces a real meal.

💡

The honest bit: This is the difference between an AI that chats and an AI that acts. MCP is what makes agents actually useful.

🐍

The Hands + Actions

Python

The chef's ability to physically do things.

+

Python is the chef's hands. It's the ability to chop ingredients, turn on the stove, mix, measure, and bake. This is execution — where ideas become real output.

Python is the most widely used programming language for AI because it's readable, powerful, and has an enormous library of tools built specifically for data and machine learning.

❌ Without Python

Just ideas and instructions. Beautiful plans. Nothing shipped.

✅ With Python

Real output happens. Files get created. Data gets processed. Systems get built.

💡

The honest bit: You don't need to write Python to use AI. But understanding what it does helps you know what's possible.

🐼

The Prep Bench

Pandas

Industrial-scale ingredient prep for bulk data.

+

Pandas is the prep station where you handle large amounts of ingredients efficiently. Chopping one onion is easy. Chopping 10,000 onions instantly? That's Pandas.

It sorts ingredients, groups them, measures precisely, and hands them to the chef ready to cook. It's basically industrial-scale meal prep for data.

❌ Without Pandas

You're manually sorting through thousands of rows of data. One. By. One.

✅ With Pandas

Instant. Sorted. Grouped. Ready to analyse. What took days takes seconds.

💡

The honest bit: Most business data problems are actually Pandas problems — messy spreadsheets, inconsistent formats, data that needs cleaning before it's useful.

How it all works together

You say: "Cook dinner
for 100 people."

Here's what actually happens behind the scenes when you give an AI agent a real task:

1
📚

RAG checks reality

What recipes work + what ingredients actually exist in your business.

2
🔌

MCP connects the tools

Access the oven, fridge, calendar, CRM — every appliance in the kitchen.

3
🐍

Python executes

The hands actually cook the food. Real output. Real files. Real results.

4
🐼

Pandas handles scale

Manages massive ingredient quantities — 10,000 rows of data, instantly.

5
🧠

AI decides + explains

The chef decides what to make and explains every decision to you.

🧠 Clean takeaway

AI

The brain

RAG

Checks reality

MCP

Tool access

Python

Does the work

Pandas

Handles scale

⚠️

The honest bit

Most people are talking to the chef but never letting them into the kitchen.

That's why results feel limited. You're getting the conversation, not the cooking. The real power comes when the chef has access to your actual kitchen — your data, your tools, your systems.

What's next for you

Want to see what your dream kitchen looks like?

This is where it gets actually powerful. Wonder Conductor is the 12-week programme where we build your full AI stack — not just the theory, but the real kitchen, wired up for your specific business and life.

Made with wonder by Sarah Pirie-Nally & Manus AI