For 13 years I ran a business that grew to £2.2 million a year, and it took a full team to keep it moving. The one I run now fits on a single screen. Most mornings I open that screen, instead of the seven tabs and three apps I used to bounce between, and it tells me who to follow up with, where every deal sits, how my recurring revenue is tracking against the target I set myself, who has been on my site, and what I should post today. By the time I have read it, I know exactly what my day is for.
I call it my Hub. It runs locally on my own machine, built with Claude Code, and it pulls my whole business into a single source of truth. A few days ago I shared the AI prompt behind the content side of it and it took off, so I recorded a full walkthrough of the thing it plugs into. Here it is.
I want this article to do two things. Show you what an AI-native business actually looks like when it is running, not in theory. And give you the thinking underneath it, because the tools will keep changing and the thinking will not.
Ninety per cent of your day should be on what makes you money
Here is the principle the whole system is built on. The vast majority of your activity in business should be pointed at the things that make you money. New leads, live conversations, sales, the work that moves revenue. Most of us do the opposite. We fill the day with busy work because busy work is comfortable and it feels like progress.
So when I open the Hub, the first thing it does is push my next action at me. The next lead, the warm follow-up, the message that should go out today, all of it backed by what the system already knows about that person from their LinkedIn, their posts, their site. I do not sit down to a blank screen and decide what matters. The system has already done that, and I get to work.
That sounds like a small thing. It is not. The blank screen is where most good intentions go to die.
One source of truth, or no truth at all
For years I ran a business where the truth lived in too many places. The pipeline was in someone’s head, the numbers were in a spreadsheet that was a week out of date, and the real status of a client sat in an email thread nobody else could see. Scattered truth like that is how you end up making big calls on a gut feel and a stale spreadsheet.
The Hub fixes that by being the one place everything reports into. My pipeline is two-way synced with my CRM, so when I move a deal on the dashboard it updates everywhere, and when activity happens in the CRM it shows up here. Right now there are 27 live prospects in it. They came out of an outreach list I built by scraping 1,868 company owners across the UK and Ireland, then working them down into real conversations. Around 90 of those are active right now and will funnel into the pipeline over the next quarter.
Then there is the money. I am a nerd about this. I have set myself a monthly recurring revenue target, before any project work goes in, and the dashboard shows me exactly where I am against it, every renewal, what each relationship is worth per month, when it comes up, and how healthy it is. I am most of the way there. The point is not the number. The point is that I can see it without asking anyone, including myself.
The signals that tell me who is warming up
Underneath all of that, the Hub watches for intent. Someone might like a couple of my posts, connect on LinkedIn, then reply to an email a fortnight later. On their own those are noise, but together they form a pattern, and the pattern is buying intent. The system surfaces those people, so I spend my outreach on the ones already leaning in rather than firing messages into the dark. That is the difference between activity and progress.
The analytics that tell me the truth, even when I do not like it
The Hub pulls my website analytics straight from PostHog. It shows me visitor trends, where people come from, and where they drop off. I rate it as highly as Google Analytics, and for funnels I rate it higher, because I can wire it into Claude through the API and build the exact views I want.
Here is a decision it helped me make. One article was driving about 70 per cent of my traffic. Plenty of people would protect a number like that with their life. I de-indexed it, on purpose, because the people it brought were not my people. They were never going to buy what I do. Traffic that does not turn into the right conversations is a vanity number, and vanity numbers cost you focus. The dip was the plan working.
The content engine, and why every word is still mine
This is the part people ask about most, so let me be precise about how it works, because the how is the whole point.
I used Apify to scrape 26 of the best accounts in and around my space, which came to 1,949 individual reels. For each one the system took the video, the caption and the audio, sent the audio to AssemblyAI to transcribe it, and built a database of what was working and what was not. Hooks, angles, structures, the frameworks behind content that performs.
That gives me the shape to work within. The words are a separate job, and they are all mine.
They have been for seven months. Every meeting I have in person goes through a transcription device. Every online meeting goes through Google Meet. I have been capturing everything I say, and from that I extract my own frameworks, my own stories, and the way I actually talk, all kept as files on my machine. So when the engine writes, it takes the proven shape from the reels and pours my real material into it. It sounds like me because it is me. If you are not recording and transcribing your own thinking yet, start today. One day you will want a seven-month head start, and you will wish you had begun the second you read this.
Watch the video and you will see the agents do this live. One scouts the competitor reels and finds the angles. One writes the hooks and scripts, and every script has to pass my voice check before it goes anywhere. One plans the calendar. One tracks how my posts perform so the next ones get better. One logs my DMs, because I will never automate outreach, but I do want every real conversation captured in the sales process. Everything routes through my approval. Nothing goes out that I have not seen.
It is not only sales and content
The same machine runs the quieter work too. I have a set of skills on my computer for the jobs I do over and over: processing a meeting into notes and frameworks, turning a podcast into content, drafting a proposal, reviewing my search performance, auditing a competitor. And one I would not give up now, a file-sorting skill that takes everything dumped on my desktop, drops it into an inbox folder, and files it where it belongs. It saves me hours every week, and those are hours I get to point back at the work that matters.
None of this is new thinking. Only the tools are new.
I want to be honest about something, because it matters. This is not me discovering automation in 2026. When I ran my agency, we built a small automation that took about 15 minutes to set up and then quietly ran across roughly 70,000 projects over the years that followed. Same idea then as now. You find the repetitive thing, you build the system once, and you let it run while you spend your own hours on the work only you can do.
What has changed is the cost and the speed. The setup that would have taken a team and a budget a few years ago, I am now running on my own, cheaply, from one machine. I grew that Belfast agency from £400,000 to £2.2 million over 13 years before I sold it, and I genuinely do not see why a solo business cannot reach a million pounds with tools like these. The gains are that real now.
What I would actually take from this
You do not have to learn every tool out there, and you should not try. The skill worth building is being able to run a few systems on your own machine. Nobody is asking you to live in the terminal all day. Being comfortable enough with code to run a few things locally is plenty, and that is about the cheapest edge in business right now. Hardly anyone is using it.
A warning to go with it. The speed AI is moving at is its own problem. Trying to keep up while you look after clients and run the business is a real cognitive load, and I will not pretend otherwise. So do not chase all of it. Invest where it makes sense for you, build the one or two systems that take the busy work off your plate, and protect your attention for the work that grows the business.
That is the whole game. Point your best hours at what makes you money, and build systems to handle the rest.
If you want an honest read on where your own business is leaking time and focus before you build anything, my Agency Health Scorecard gives you the answer in about two minutes.