Do I need to learn any of this?
No. That is the point of having it built for you. You describe how the work actually runs, and it is built to match. You never need to know what any of it is called.
What can you build?
Twelve pieces in four groups: private AI (a private AI server and a company brain), workflows (paperwork, portal, revival and deadline agents), custom code (connecting code, spreadsheet to app and an owner's dashboard), and safety and people (an AI safety review, team training and monthly care). Each can be built alone; together they make one system.
Can the AI run on our own computer?
Yes. A private AI server runs a model on a machine in your office, so client data stays in the building. Your Blueprint says which work runs there and which uses outside services you approve.
Do we have to switch software?
Usually no. I build inside what you already use: email (Gmail or Outlook), text messages, calendar, your CRM and your books (QuickBooks and the like). If a tool can't do the job, the Blueprint says so before you spend anything.
Does it answer phone calls?
It works written messages: texts, web forms, email and lead-site messages. Calls stay with your team, and your Blueprint says whether a call line belongs in the plan.
How do I keep the cost of running AI down?
Three habits keep it low. Every call to a model is written down with what it cost. Each task runs on the smallest model that still meets your quality floor. What repeats is served from memory instead of processed again from the start. On my own systems, read from September 8 to October 9, 2026, one long working session served hundreds of millions of tokens that way. Before you commit to anything, you read what one result costs on your own numbers.
How is a build priced?
Builds are five-figure engagements. Your Blueprint names the exact investment, in writing, before any work starts. The Blueprint itself is free.
Can I skip the Blueprint and just hire you?
Yes. Send the application and say so in the last box. If your answers fit, you pick a call time on the spot and we start there. The Blueprint still gets drawn, because the build needs a scope, and it becomes the document we both work from.
How long does a build take?
Your Blueprint sets the timeline. Much of it is inspection, because the system works beside your team on real work before it works alone.
What if it makes a mistake?
It asks before anything important goes out, and it logs every action. During inspection it drafts only. When it is unsure, a person decides.
Whose data is it?
Yours. The system runs inside your accounts, and client information goes only to tools you approve.
Is it safe for my staff to use public AI tools with client information?
Not without rules. An AI safety review finds where staff already paste client data into public AI tools, then sets the rules and a safe setup, which can include a private AI server in your office so client data stays in the building. Team training puts the safe habits into the daily work.
Is it legal to let AI answer my customers?
Yes, with rules, and every build follows them. In California, an automated system that talks to customers to make a sale has to say it is automated. Marketing texts need the customer's consent, and marketing emails need an opt-out. Your Blueprint names the rules that apply to your business.
What kinds of companies do you work with?
Owner-led companies of 5 to 100 people: trades, property managers, offices and suppliers. They all run on the same things: calls, email, quotes, schedules and follow-up.
What does governed mean?
The rules are written down before the build. Every action is logged where you can read it. Anything irreversible waits for a person, and every part has an off switch you hold.
Why a custom build instead of an app?
Because your company does not run like anyone else's. I draw the plan first, prove it on your real work, and care for it after launch.