AI Terms
A glossary for people who'd
rather be running tours.
AI terminology explained without the jargon or the assumption that you have a computer science degree. A crash course from zero to dangerous.
The Basics
AI (Artificial Intelligence)
Software that can learn patterns, make decisions, and generate content without being explicitly programmed for every scenario. Think of it as a very fast intern who's read everything on the internet but has zero common sense.
Why it matters
AI can draft your emails, analyse your reviews, personalise your marketing, and automate the admin that's eating your evenings. It won't replace you. It will replace the version of you that spends three hours on a spreadsheet.
LLM (Large Language Model)
The engine behind tools like ChatGPT and Claude. It's been trained on billions of words and can generate human-sounding text, answer questions, and follow instructions. It doesn't "understand" anything. It's spectacularly good at predicting what word comes next.
Why it matters
This is the technology that lets you have a conversation with a computer and get something useful back. Every AI tool you'll use in the next five years is built on one of these.
Prompt
The instruction you give an AI. "Write me a reply to this guest complaint" is a prompt. The quality of what you get back is directly proportional to the quality of what you put in. Garbage in, garbage out — but politely.
Why it matters
Learning to write good prompts is the single fastest way to get value from AI. No code, no technical skills, just clear thinking and specific instructions.
Token
The unit AI uses to process text. Roughly, one token equals about three-quarters of a word. Every time you use an AI tool, you're spending tokens — both on what you send and what it sends back.
Why it matters
Tokens cost money. Understanding them helps you understand why AI tools have usage limits and why a 10,000-word document costs more to process than a quick question.
Context Window
The amount of text an AI can consider at once. Think of it as the AI's working memory. A small context window means it forgets what you said five minutes ago. A large one means you can feed it your entire operations manual and ask questions about it.
Why it matters
This determines how much of your business context you can give the AI in a single conversation. Bigger context windows mean smarter, more relevant answers.
Working With AI
Vibe Coding
Building software by describing what you want in plain English and letting AI write the code. You're the architect with the vision. The AI is the builder with the tools. No programming knowledge required — just the ability to articulate what you need.
Why it matters
This is how non-technical founders are building custom tools, websites, and automations that would have cost £30k from an agency eighteen months ago. Your domain expertise IS the skill.
Hallucination
When AI confidently generates something that's completely wrong. It doesn't know it's wrong. It will present fabricated facts with the same conviction as real ones. It's not lying — it genuinely can't tell the difference.
Why it matters
Never trust AI output without checking it. This is especially critical for anything customer-facing, anything involving prices or dates, and anything you'd be embarrassed to get wrong in front of a guest.
Agentic
AI that can take actions, not just generate text. An agentic system might read your emails, check your calendar, draft a reply, and send it — all without you lifting a finger. It's AI with hands, not just a mouth.
Why it matters
This is where AI goes from "helpful tool" to "actual team member". Agentic systems can handle multi-step workflows that currently require a human sitting there clicking buttons.
RAG (Retrieval-Augmented Generation)
A technique where AI searches through your specific documents before answering a question, rather than relying purely on its training data. You give it your SOPs, your FAQs, your product descriptions — and it uses those to answer accurately.
Why it matters
This is how you make AI that actually knows YOUR business, not just generic information. Want a chatbot that can answer questions about your specific tours? RAG is how.
Fine-tuning
Training an AI model on your specific data so it learns your patterns, your tone, your way of doing things. It's like hiring someone and spending six months getting them up to speed — except it takes hours and they never forget.
Why it matters
For most operators, you don't need this yet. Prompts and RAG will get you 90% of the way. But as AI becomes more embedded in your operations, fine-tuning is how you get that last 10% of accuracy and consistency.
Markdown
A simple way of formatting text using plain characters. Headings get a # symbol, bold text gets wrapped in **asterisks**, lists use dashes. It looks like a slightly decorated text file. The reason it matters: every major AI tool reads and writes Markdown natively. When you structure your instructions, specifications, or business documentation in Markdown, AI understands the hierarchy instantly. No special software needed — any text editor works.
Why it matters
If you write specifications, briefs, or documentation for AI tools, Markdown is the lingua franca. Learn the basics (headings, bullets, bold) and your AI interactions become dramatically clearer.
Building Things
API (Application Programming Interface)
A way for two pieces of software to talk to each other. When your booking system sends a confirmation email, that's an API at work. When Zapier connects your form to your spreadsheet, APIs are doing the heavy lifting behind the scenes.
Why it matters
APIs are the plumbing that makes automation possible. You don't need to build them, but understanding that they exist helps you understand what's possible when someone says "we can connect that".
The Articulation Method
Jack's approach to building with AI — the idea that the most valuable skill isn't coding, it's the ability to clearly describe what you want built. Domain expertise plus clear articulation equals functional software, no programming degree required.
Why it matters
If you can explain how your business works to a new employee, you can explain it to an AI. That's the entire skill. Everything else is tooling.
Specification Engineering
The practice of writing detailed, structured descriptions of what you want built — clear enough that an AI can execute on them without constant hand-holding. It's the deepest layer of working with AI, beyond prompts, context, and intent.
Why it matters
This is the difference between "build me a website" (vague, expensive, disappointing) and a 20-page spec that an AI can turn into a working product in a weekend. The spec IS the skill.
Stack
The combination of tools and technologies that make something work. A website stack might be Next.js (the framework), Vercel (the hosting), Sanity (the content management), and Supabase (the database). When someone asks "what's your stack?" they're asking which tools you chose and how they fit together. Getting the stack right matters because these tools need to talk to each other. A good stack is one where the pieces connect cleanly and you're not fighting the tools.
Why it matters
Understanding your stack helps you make better decisions about what to build, what to buy, and where AI fits in. Every new tool needs to integrate with what you already have.
The Money Bit
AI Readiness
A measure of how prepared your business is to benefit from AI. It covers your digital infrastructure, your data quality, your team's comfort with technology, and your operational workflows. It's not about being "high-tech" — it's about being ready.
Why it matters
Take the AI Readiness Score to find out where you stand. Knowing your starting point prevents you from wasting money on AI tools you're not ready for — and highlights the quick wins you're missing.
Automation
Using software to perform repetitive tasks without human intervention. Every time you set up an "if this, then that" rule — like auto-sending a review request after a tour — that's automation. AI makes automation smarter by handling tasks that previously needed human judgement.
Why it matters
The average tour operator spends 15+ hours a week on tasks that could be automated. That's two full working days you could spend on the parts of the business that actually need a human — like talking to guests and creating exceptional experiences.
Model Selection
Choosing the right AI model for each task. Claude Opus is brilliant at complex analysis but costs more per query. A smaller model handles routine classification for a fraction of the price. The same principle as hiring: you don't bring in a senior consultant to sort the post. Tracking which model handles which task — and what each one costs — is how you keep AI economically viable as you scale usage across your operation.
Why it matters
AI costs are directly tied to model choice. Operators who match the right model to each task spend a fraction of those who default to the most powerful option for everything. This is the difference between AI as an experiment and AI as infrastructure.
Now what?
Knowing the words is step one. Now put them to work.