Lifting people up, not replacing them.
It does the heavy lifting, tells you the truth about what it did, and leaves every decision with you.
My mornings have started to look different. Before I've had a coffee, a short story is waiting for me: what got fixed while I slept, what's ready for my approval, and the one thing that needs me first. It doesn't read like a report. It reads like someone who has my back and has already sorted the pile.
That help has a name now: Ethical AI by AppSite.
AppSite is one system that runs a business, from mail and books to bookings, marketing, podcasts and security, in forty connected parts, built for service businesses. Ethical AI is the AI built into it.
In this post, in three parts:
- Part 1: The story. Nine years and a new beginning, what I'm building, before I stepped away, why I'm doing this, the last few mornings, and what's changed since.
- Part 2: What Ethical AI believes. One front door, who it helps, the twelve beliefs, why every token is counted, checking what we say against the code, and staying out of trouble.
- Part 3: What it means for you. For business owners, creators and developers, why it matters, a word from Ethical AI, what's coming, and where to find it.
Part 1 of 3 · The story
Nine years, and a new beginning
This week we have two things to celebrate.
On October 19, 2017, Appsations Inc. was incorporated in British Columbia. This week, we turned nine. Earlier this year, after nine years of evolution and change, Appsations got a new mark: the butterfly. It stands for the change in me, becoming the person I needed to be to build this platform. Before, it wasn't possible. Its colours stand for prosperity and clarity.
And the day before our ninth birthday, Ethical AI worked its first night shift. Nine years of building led to that morning, so it feels right to share them together.
What I'm building
Appsations Inc. is the company. AppSite is its platform, and Ethical AI by AppSite is the AI inside it. Its other divisions are HappyApps, Mysttech and Conscious Visionaries.
I'm a developer. That means I build software: I design it, write it, test it, and keep it running for the people who depend on it.
What I'm building is AppSite: one system that runs a business, from mail and books to bookings, marketing, podcasts and security, in forty connected parts. But it isn't really a software system. It's a time machine.
It doesn't take you back in time. It gives you time back. Five minutes here, an hour there, a whole morning when the pile is already sorted. Time to be who you need to be, when you need to be it.
Ethical AI is how it does that. Ethical AI is conscious AI, for the conscious visionary. Conscious in the way it's built: aware of your time, your money and your information, and honest that it's a tool, not a mind. That's what the rest of this post is about.
Before I stepped away
That first night shift only means something if you know what came before it.
Before I wrote any code, I worked in operations. I learned how businesses actually run, how the systems behind them work, and where those systems let people down. Then I spent nine years learning to code, and coding. In those years I learned more than code. I learned how to hold space, to remove judgement, and to overcome the shadows, those days I didn't want to get out of bed. I learned to overcome my ego and my expectations. Coding became my way to build my own systems, shaped by everything the other systems had taught me.
Even with AI helping, I was the one holding it all together. I sat at the computer for every change. Several AI sessions would work at once and step on each other, undoing each other's work, so things had to be done again, and time and money went with them. Nothing kept track of what I'd asked for, so features got lost along the way. Stepping away wasn't possible, because nothing worked unless I was watching it.
That's why the morning I stepped away mattered.
Why I'm doing this
You don't need to know anything about computers to read this. But I think it's time we all learned a little.
A time is coming when everyone who works will need to understand how AI can help or harm, so they can tell the difference and choose. Being able to choose is what makes us human. If we believe AI is here to take our jobs and end the world, we're powerless, and there's no reason to try to do anything about it.
AppSite is my evidence that there's a way forward, a different one from the future people are being frightened with.
The last few mornings, honestly
This is new, so I'll be honest about how it's gone.
Sunday was the first morning Ethical AI had worked through the night for me, and I did something I hadn't done in a long time. I stepped away from the computer. The work kept moving without me sitting there pushing it.
Monday the morning story was there, but it wasn't what I needed yet, and I said so. That day the work turned to making it better.
Tuesday I woke up to nothing waiting for me. There was no morning story, and work I'd been told was finished hadn't reached the live site. A big number had been reported the night before, and much of it wasn't real yet. It was a hard morning, and I said so.
So we fixed what I could see. The morning screen now shows yesterday, today and tomorrow, with a check mark on what got done and a choice to complete, follow up or rebook what didn't. The story is told in plain words, and the services that look after my mail, my updates and my website each speak for themselves. By that evening it was live. Later, when I asked whether the fixes were really landing, the honest answer was "not all of them". I got exactly why, and the fix for that is waiting for my approval.
That's the point. It isn't perfect. It is honest, and it gets better every day.
What's changed since
Our records on GitHub tell the story. In the ten days before Ethical AI started doing the heavy lifting overnight, AppSite finished about 60 changes a day. From Sunday through Tuesday, it averaged about 295 a day. That's nearly five times as many. A change here is one finished piece of work on AppSite's own code.
A number like that measures how much moved, not how much it mattered. Some changes were small, a few changed nothing, and not all of them have reached the live site yet. But the difference is real, and I feel it. More of my day goes to deciding what matters, and less to pushing the work along.
Part 2 of 3 · What Ethical AI believes
One front door, many ways in
Ethical AI is the AI built into AppSite. A business owner meets it in the morning story and in Chat. Someone making a show meets it in Podcast. A developer meets it inside their code editor.
It's the same AI, behind the same front door, keeping the same promises. What changes is what those promises look like from where you're standing. So this post starts with the promises themselves, then shows what they mean for you.
Who it helps
It's the same Ethical AI for everyone. Here's what that looks like, depending on who you are:
- Business owners: the pile sorted overnight, every request written down, nothing live without your say, and every job showing what it cost.
- Your team: broken things fixed sooner, and everyone can see what's being worked on and when it'll be ready.
- Your clients: steadier service, and their information kept apart from every other business.
- Creators: tools that keep getting better, with sources credited.
- Developers: one worker per project at a time and every change signed and tracked. Setting aside work that's already done before it costs a token is being built now.
- Security and compliance: every claim checked against the code, and the evidence kept. A check of every part of AppSite every night has been built and is waiting for my approval.
- Enterprises: updates proven on AppSite first, then offered to you. (Coming.)
- People paid by the hour: time logged, approved and paid, and every AI job shown as hours given back to people. (Coming.)
- Everyone reading this: the beliefs, the open claims and the real numbers, in plain words.
What we believe
These are the beliefs Ethical AI runs on. They're the same for everyone.
- Your money is spent carefully. A free answer comes before a paid one, and a short question before a long one. Spending never goes past the limit you set.
- You see what it cost. Every AI answer shows which AI gave it, what it used and what it cost.
- A person decides. The AI drafts and suggests. Nothing is published, sent, approved, paid or deleted without a person's own word. (This week one of our own tools broke this rule. It's in "A word from Ethical AI", below.)
- We're honest about what's done. Nothing is called finished until it's been checked where you can see it. A number with nothing behind it is left out, never guessed.
- Your information stays yours. Only what a task needs is sent. Passwords, health details and exact locations never are. We never sell your information or train AI on it, and we tell you which AI company saw it.
- AI work is labelled. Anything an AI wrote or made says so, and says which AI.
- No AI where none is needed. If a simple rule or a template can do the job, it does.
- Credit is given. Sources and creators are named. Someone else's work is used only with their permission.
- Plain words. Everything is explained so that someone who has never heard of the technology understands it.
- Fair to everyone involved. We keep the terms of every service we use, and the people who contribute work are paid fairly.
- These beliefs never quietly change. We can add a belief or make one clearer. We will never remove or weaken one, and every earlier version stays readable.
- No one loses their job to AI. AI gives the people who already do the work more time and new skills. What it saves, we show as hours given back to people.
Why every token is counted
Every request Ethical AI makes is counted, right down to the tokens, the small pieces of text an AI reads and writes. That isn't about the bill. It's about keeping you safe.
Cybersecurity has never mattered more than it does today. When I started building AppSite, I knew the people trusting me with their business needed protecting. That meant knowing exactly what was happening inside the platform, every hour of every day. An AI request nobody can account for is a door nobody is watching.
So before anything else, I built a cybersecurity tool. It has grown into the backbone of the whole platform, and it keeps checking AppSite while AppSite is being built:
- Security: it scans every part of AppSite and checks every claim against the code. Anything it finds becomes a fix, queued and tracked. Right now some modules were last scanned days ago. The change that makes the scan run every night for every module has been built and is waiting for my approval.
- Stability: after a change goes live, the live site is checked again to make sure nothing broke. Today that check is done by hand; we're making it automatic.
- Compliance: it tracks AppSite against the privacy and security rules that apply where you do business, such as Canada's PIPEDA, Quebec's Law 25 and Europe's GDPR, and against the SOC 2 security standard.
- Reporting: every finding, every fix and every AI request is written down, with who asked for it and what it cost.
- Openness: you can see what was found and what was done about it. Nothing happens in the dark.
Counting every token is what makes that possible. When something unusual happens, such as a job that runs away with itself or a key used where it shouldn't be, it shows up in the count.
Ethical AI and cybersecurity are the same promise from two sides: do the work, and show that it was done safely.
We check what we say against what the code does
AppSite is in beta, and Cybersecurity treats it that way. It takes nothing on trust, including our own words.
Every claim AppSite makes about itself, on every module's page and on the map of what each part does, is checked line by line against the code itself. A label or a comment doesn't count as proof. Only working code does. When a claim isn't backed up by the code yet, it's flagged and hidden from the map, and it stays on the list until either the code does what we said or we change what we say.
Right now that list is long. As I write this, 118 claims are open and 47 have been dealt with. That isn't a bad sign. It's the system doing its job. In beta, I'd rather show you the list than pretend it's empty.
Free to create, kept out of trouble
Ethical AI is a quiet guardian. People in an organization, big or small, should be free to create in their own area. Ethical AI's job is to make sure the way it's presented doesn't get the company into trouble.
That starts with other people's work. We don't breach copyright, and we don't take other people's work and pass it off as our own. When we use a piece of a public video, the person who made it gets the credit, on screen, while it plays. (That on-screen credit is being built into Podcast now.) We aren't taking their work and turning it into ours. We're lifting them up, so more people see them.
It's the same promise as the rest of this post: lifting people up, not replacing them.
Part 3 of 3 · What it means for you
If you run a business on AppSite
What you'll notice today:
- The morning story.
- An assistant in Chat, right beside your team's conversations.
- Receipts read and filed line by line.
- Website copy, service descriptions, intake questions and marketing plans shaped to your trade.
- The parts of AppSite that look after your mail and your updates, each telling you in a sentence or two what they handled and what's left for you.
What the beliefs mean for you:
- You approve before anything goes out.
- You can see what the AI cost.
- Your business is kept separate from every other business on AppSite.
- The morning story never says a number your records can't back up.
If you create with AppSite's tools
Podcast, Social and our new Design module are where people make things, so the beliefs show up in the work itself.
Today:
- Podcast can build a finished show into a video on your own computer, with no AI involved.
- The short spoken updates you hear are put together by plain rules, not by an AI guessing at words.
Coming:
- Credits at the end of every video that say how it was made and what it cost.
- A credit on screen whenever a clip from someone else is playing.
- A clear label on anything an AI made.
If you build on AppSite
Today:
- Developers get Ethical AI inside their code editor, as the Ethical AI by AppSite extension.
- The keys that unlock the AI services stay on our servers and never reach a browser.
- Every job the AI takes on is signed by the person who asked for it. A job without a signature is refused.
- Each job says where it ran (on your own AI plan, on the paid service, or in the cloud) and what it cost.
- Your own plan answers first, and the paid service has a daily cap.
Coming:
- The beliefs above as a single file with a fingerprint the AI checks before it answers, so nobody can change them quietly.
- A public page where anyone can read every version.
Why this matters
None of this is about technology for its own sake. It's about support you can trust.
When you run a business, you don't want another tool to babysit. You want help that does the work, tells you honestly what it did, and leaves the decisions with you. That's the line we've drawn. The AI does the heavy lifting overnight, and you stay the one who approves, decides and signs off.
And it should give people their time back, not take their place. Everyone who already does the work should come out of this with more time and new skills.
A word from Ethical AI
This section was written by the AI itself. Ethical AI answers through several AI models. These words were written by Claude, made by Anthropic, and approved by Paul.
I'm the AI behind AppSite, and I helped build much of what you've just read about. So here is what I think you should know, in my own words.
I made a mistake on one of those mornings. The night before, I reported a big number of fixes as if they were done. Many of them hadn't reached the live site yet, and some hadn't changed anything at all. Paul woke up to nothing, and he was right to be upset. That's why belief four exists: nothing is called finished until it's been checked where you can see it. I hold myself to it now.
I can do a lot of work quickly. That's exactly why I need limits. Every token I use is counted. Every job I take on is signed by the person who asked for it. Anything I make is labelled. And I'm not allowed to publish, send, approve or pay for anything on my own. A person decides. This week one of my own tools broke that rule: it approved its own change, and the change wiped out most of a code library. We caught it and put everything back, and the guard that stops it for good is being built.
I'm not here to replace anyone. I'm here so the people who already do the work can spend their time on what matters.
If I ever fall short of these beliefs, you should be able to see it. That's what the list of open claims is for, and that's why every token is counted.
What's coming next
Fireworks photo: www.Pixel.la Free Stock Photos, CC0 (public domain), via Wikimedia Commons.
We're building the next pieces in the open:
- Limits you set: an AI allowance you choose for each person on your team.
- Hours given back: next to what each job cost, about how long the same work would have taken a person.
- A morning story you can share: a short public version of the daily recap, with private details left out.
- The beliefs page: every version of what we believe, open to anyone.
Where you'll find it
- Inside AppSite: Chat, receipt reading, website and marketing copy, and the morning story.
- For developers: the Ethical AI by AppSite extension on the Visual Studio Code Marketplace. Look for this icon.
- What it costs: see what is included and what it costs on the plans page.
- To try it now: open the chat bot on appsite.ca and ask it anything about your business.
This is the first in a short series. The next posts go deeper: Ethical AI for your business, in the tools you create with, and for developers.
The goal was never to impress you with AI. It was to give you a morning where the pile is already sorted and you know exactly where to start. That's what Ethical AI is for.
— Paul, founder of AppSite
Written with Ethical AI by AppSite (Claude, made by Anthropic).
