11 AI buzzwords demystified

Below we've explained the terms our teams are most often asked about:

1. AI (Artificial Intelligence)

Computers doing tasks that usually need human thinking, like recognising patterns, answering questions, or making suggestions. AI has been around since the 1940s in various forms.

2. Artificial General Intelligence (AGI)

A hypothetical future AI that can learn, reason, and perform most tasks at a skilled human level across many domains; not just a single niche. Big tech is pursuing this with record investment because it could reshape jobs and productivity. In our view, it’s still some distance away and overhyped; access to high‑quality data remains a major bottleneck.

3. Prompt

What you type or say to an AI to get it to do something. Think of it as your instruction or question. A good prompt provides context; sets expectations and guidelines for the output you want.

4. LLM (Large Language Model)

A type of AI trained on huge amounts of text so it can predict words and generate human‑like writing, answer questions, and follow instructions. Examples of LLMs are ChatGPT or Claude. These AI models “feel” human which has led to the recent excitement around AI.

5. Token

A small chunk of text (often a piece of a word) that AI models read and write. Models count input and output in tokens, not characters or words. With many AI models you pay by tokens, as a rule of thumb 150 words is roughly around 200 tokens which would cost about £0.00020.

6. Agentic AI

AI that can break a goal into smaller steps and carry them out on its own, like looking things up online, pulling info from your business systems (e.g., CRM or finance software), and taking actions on your behalf without much guidance.

7. Hallucinations

When an AI confidently makes things up, plausible‑sounding but factually wrong answers. This happens because AI isn’t really thinking, it’s pretending to think by predicting the “best” sentence to respond to you. You can manage hallucinations by grounding AI responses with data, e.g. from your own documents.

8. GenAI (Generative AI)

AI that produces new content, text, images, audio, code, and more, rather than just analysing what already exists. It’s the broader category that includes LLMs, all focused on generating new material.

9. Microsoft Copilot

An AI productivity assistant that’s built into tools you already use like Word, Outlook and Teams. Under the hood Copilot uses ChatGPT AI connected to your data securely in the Microsoft 365 ecosystem.

10. ChatGPT

A widely used conversational AI by OpenAI that answers questions, drafts text and helps with tasks via chat. ChatGPT’s launch is the thing that generated this new wave of excitement and OpenAI are one of the market leaders in the Generative AI space.

11. Claude

A conversational AI from Anthropic, designed to be helpful and careful, widely used for analysis, writing, and coding support. It’s a strong alternative to OpenAI for certain tasks, especially analytical work. Anthropic is a market leader with a strong emphasis on responsible AI.

Get in touch

If you’d like a straight‑talking, honest conversation about AI and what it can and can’t do, get in touch at ai@waterstons.com .

To discover more ways to mind the hype and navigate your way through AI, click here.

You'll find more great insights, upcoming webinars, and clear, actionable advice and guidance from our experts.