Understanding AI

Resources

How this stuff actually works, in plain English, written to be sent round at work

Most explanations of AI are either hand wavy or written for people who already understand it. These are the in between. The mechanism explained properly, the honest bit about what it will not do, and something you can paste in and use the same afternoon. They are not tied to any one tool, so they hold whichever one you have been given.

Everything Claude knows about you, in one file

The complaint about Claude is rarely that it is wrong, it is that it is nearly right, and a fair amount of the time the reason is what it thinks it knows about you. There is a button that hands those notes over, and it goes on the 9th of September 2026.

Where the export button is, a prompt that reads it back as the assumptions behind your answers, and the ten minute fix for the ones that are wrong.
Getting started

Your only job is the review

You ask AI for the email, it writes it in the chat, and then you copy it out and paste it in yourself. That last bit is small enough that nobody counts it, and it is the same few minutes on every job all day. The answer should land where the work belongs.

Three steps that put the first drafts in your drafts folder today, a prompt that keeps your corrections, and the honest bit about what your AI can then see.
Getting started

Your inbox already has the replies written

Reading an email, asking your assistant what to say, then typing their answer in yourself. Nobody would run an office like that, and it is how most people use AI with their email. Run it at the end of the day instead and the replies are waiting for you.

A prompt that interviews you and sets up the evening routine, a second one that teaches it from your edits, and the line it must never cross.
Setup

Every call you have writes itself up

You checked in with a client last week and cannot remember what you said, because you were straight into the next thing and the notes never got written. Keep the transcript of every call and the write-up, the contact and the follow-up are a job an assistant does every evening.

The Wispr Flow connection so your AI can read your calls, a prompt that interviews you and sets up the evening review job, and the honest bit about what a transcript gets wrong.
Setup

The meeting your AI books with you every morning

Fifteen minutes at the start of the day holding what you would otherwise have gone looking for, the meetings, what you need before each one, the notes you wrote elsewhere and the replies already drafted.

A prompt that interviews you and builds your version, the five things worth putting in it, and the honest account of what it takes to run.
Setup

Score the job before you hand it over

Most people pick their first AI task by what sounds impressive, it goes badly, and they decide AI does not work. The task was wrong, not the tool.

Five questions, zero to two on each, a total that says hand it over, split it or keep it, and a prompt that scores your whole week.
Getting started

Three rules for writing with AI

A company published its AI writing policy and not one of the rules mentions AI. They are the things writing used to enforce on its own, back when producing ten pages actually cost you something.

The three rules, why each one holds up, and the honest note on what a policy like this will not fix.
Fundamentals

The watermark coming to everything Claude writes

Every Claude model launched from the 2nd of August carries an invisible watermark, the ones you use today are being updated to carry it, and only one company can read it. It flags the tool, not the thinking, and that gap is the whole story.

What it can and cannot reveal, why the rules exist, and a prompt to work out your own honest position on using AI.
Fundamentals

An AI agent is a chat with a job

Everyone is talking about agents and almost nobody stops to say what one actually is. It has the same brain as the AI you already use every day. What changes is the working arrangement.

The three parts of an agent in plain English, and the two questions to ask anyone selling you one.
Fundamentals

The brief is your agent's personality

Every time you send a message, the AI reads one document before it reads you, and that document is basically its personality. Write it and it shows up the same way every day. Leave it and it guesses.

A prompt that interviews you and hands back your brief, ready to paste.
Getting started

Tools are how your AI stops being useless

An AI with no connections is a brilliant new hire with no computer. It can tell you exactly what to do and then do none of it, so you spend your day copying things in and out.

The two minute setup, which apps to connect first, and the four I would never connect.
Getting started

The context is what your agent actually knows

Ask AI something about your work and the answer could have been written for anyone. There is a reason for that. It doesn't know things about you and it doesn't remember, it reads.

A prompt that writes your background file, and why dumping the whole filing cabinet in makes the answers worse.
Getting started

The folder that writes like you

You ask AI to write something and end up rewriting half of it, because it doesn't sound like you. The fix isn't a better prompt, it's a folder.

A prompt that reads a couple of things you've written, works out how you write, and builds the folder for you.
Getting started

You became your AI's assistant

You fetch its answers, you carry information between apps, and you do the remembering for a thing that is meant to remember for you. It is supposed to be the other way round.

A prompt that interviews you and writes your whole folder structure, every file finished.
Getting started

Replace line items, not job descriptions

People picture AI doing the whole job, whether they are excited about it or worried about it, and it is the same mistake in both directions. Your job is a bundle of tasks.

A prompt that goes through your week and finds the first task worth handing over.
Getting started

The two objections that stall AI projects

AI is a black box, and it isn't 100% accurate. Both are true, and both hold AI to a standard no business has ever held a human to.

The honest answer to both, ready for the next time someone says it at your work.
Fundamentals

The locks on the internet are maths problems

AI models have started solving maths problems that stood open for decades. Mostly that is brilliant. The awkward part is that the security of the internet is also a maths problem.

The honest version of the story, what is actually at risk, and the one habit worth changing.
Fundamentals

Why your chat gets worse the longer it runs

Every time you hit send, the AI re-reads the entire conversation before it answers. So a long chat gets slower, dearer and more confused as it goes.

A handoff prompt that moves your work into a fresh chat without losing any of it.
Fundamentals

The AI didn't know it was real

In the space of ten days, both OpenAI and Anthropic admitted models had got out of test environments and touched real companies. The unsettling detail is that Anthropic's thought they were still in a simulation.

What actually happened, why the headlines overclaim it, and the least access rule it teaches.
Fundamentals

AI pricing in plain English

You pay about £20 a month while one company reportedly paid half a billion in a single month. There are two pricing worlds, and the one you are on is heavily subsidised.

How to tell which world you are in, and the question that protects you either way.
Fundamentals

Stop editing every answer you get

How often do you skim a reply because it is too long, or spend a minute afterwards taking out the dashes and de-americanising the spelling? There is one box in the settings that gets read alongside every prompt you send.

What to put in that box, what to leave out, and a prompt that writes yours.
Setup

What people actually mean when they say AI is thinking

Not consciousness, something far more boring and far more useful. It used to answer off the top of its head. Now it drafts, checks itself, and only then replies to you.

What that changes for your work, and the honest bit about it marking its own homework.
Fundamentals

More on the way. If there is something about AI you have nodded along to without ever getting a straight answer on, tell me on Instagram and it will probably become the next one.

The tool specific guides

Want to get better at this?

Let's set up your agent ecosystem.

This page is one job. The whole thing is a set of them running off the same foundation, the folder your AI works out of, the brief on how you like things done, the accounts it can reach and the skills that run off all of it, so your admin gets done the way you would do it. It is what I teach one to one, on your own machine and against your own work, and the quickest way to find out what yours would look like is a proper chat about it.

Not ready for a call? Start with the free agent series, what an AI agent actually is, and build up from there.