How We Run Caffeine
We connect the systems a business runs on and put AI to work inside them. We did our own first, and this is what it looks like on an ordinary day, including the parts that have gone wrong.
We are a small software company. For most of thirteen years, running the business ran through one or two people: support tickets, DNS changes, proposals, the newsletter, the website, the hosting. The team builds the client work. Everything around it was owner work.
That stopped being true this summer, and not because we hired anyone. We gave an AI real access to our systems, wrote down how we do things, and set it up so it does the routine work and we approve it.
It is the same setup we install for clients.
What it is connected to
It talks to each of these through its API. It is not clicking around in a browser pretending to be one of us.
- Our helpdesk
- Our issue tracker and code hosting
- DNS
- Transactional email
- The marketing contact database
- The form pipeline on this website
- Our servers
- Our YouTube channel
- Our inboxes
- Our calendars
What a support ticket looks like now
-
1
Ticket arrives
A customer emails our help address.
-
2
It does the work
Reads the ticket, checks the site if it needs to, drafts a reply in our voice.
-
3
It waits
Nothing is sent. The draft sits until a person reads it.
-
4
We approve
Usually on a phone. Say yes, or change a line.
-
5
Sent and closed
It sends the reply and closes the ticket.
Steps three and four are the whole design. Everything before them is speed; everything after them only happens because a person said so.
Early on it signed off wrong. We told it once. That correction is a rule it reads before every reply now, and it has not happened again.
That is the whole of what people mean by a system that learns. Something got written down, and something reads it every time.
What a site change looks like now
We moved this website off WordPress to a static build we host ourselves. The AI did most of the porting and most of the checking. It found videos sitting on the old server that we had forgotten about, put them on our YouTube channel, and updated the pages that referenced them. Nobody opened YouTube.
First commit to live site was under a week. Then it wrote down how it had done it, which is why the next one will be faster.
What content looks like now
Every morning we get a short list of post ideas, pulled from our inboxes and from whatever we worked on the day before. We pick one. It drafts the post in our voice. Nothing publishes until we say so. The newsletter works the same way.
What it knows
About twenty documents describe how we work: our voice, our servers, how to reply to a client, how to migrate a site, how to qualify an RFP, what we have and have not done. The AI reads the ones it needs for the job in front of it. When something changes, or when we correct it, the document changes.
What it does not do
- It does not send, publish, or change anything that matters without one of us approving it.
- It does not hold passwords for things that need a person to log in.
- It runs on one small server, alongside our automation, and it depends on one AI vendor. We would rather say that than imply otherwise.
- It gets things wrong. The setup is built so that wrong shows up as a draft nobody approved, rather than as an email a client received.
Why we now build this for other people
We did not build it to sell it. We built it because we were the bottleneck. It turned out the parts are the same for any small business: connect the tools, write down how you work, let the AI do the routine part, and keep a person on the approvals.
So it became one of the things we do, alongside the websites and the apps. It is not a separate business and it is not a product with a name. It is the same work we have always done, connecting systems, with a newer and much faster tool in the middle of it.
Want to see it?
We are happy to walk through any of this with you, including the parts that have gone wrong. Or tell us about your business and we will point at where the bottleneck actually is.
Start With an Audit