What we do
We help organizations put AI to work carefully. That means clear expectations for staff, a sound plan for what to build, and an experienced developer for the systems that need work. You work with John directly, from the first conversation to the final report.
Help your staff use AI tools well
Most organizations don't have an adoption problem. They have an expectations problem: nobody is sure what's allowed, what "done" means, or who's accountable. We fix that with a short session for mixed audiences and a one-page AI Use Standard tuned to your existing policies, which your team can adopt when you're ready.
You keep: the AI Use Standard, a Quality Procedures checklist, and the session materials.
Plan AI work that fits your budget and your data
Which tasks AI should take on, which model is right for each, and what it will cost every month. Cost is one reason to choose a model. Where your data goes is often the bigger one. Some information can go to a commercial AI service under the right contract terms. Some shouldn't leave your organization at all, and for that a model can run on your own hardware. We sort your work into those categories, choose models to match, and test cheaper or local models before trusting them. Then we set up metering so the monthly bill is predictable. Our development process and tools work the same way with any of these models, so the choice stays yours, and you can change models when prices, terms or your rules change.
You keep: a written plan with each task, where its data may go, the model chosen and why, and the expected monthly cost.
Give your developers gates, not guesswork
AI coding agents produce more code than any person can review. The answer is a written process (plan approval, tests shown, a second review, a person at the merge) plus skills that travel between projects. We bring a starter kit and leave your team running it.
You keep: a starter skill library and an instructions-file template, installed in your own repositories.
Retire the technical debt you've been living with
Old applications, undocumented scripts, the system only one person understands. An experienced developer rebuilds and documents them, with AI help that lowers the cost of work that used to be too expensive to start. The work happens under your review and on your repositories. We can also add AI features to applications you already run.
You keep: the code, the documentation and the history, all in your own repositories.
Close the holes before someone else finds them
Attackers now use AI tools to find and exploit weaknesses in software faster than ever, and old, unmaintained applications are the easiest targets. The same tools help on defense. We find and fix vulnerabilities in your applications and the libraries they depend on, bring outdated frameworks and language versions up to date, and set up automated checks that warn you when a new vulnerability affects your code. Then the next announcement means a routine update, not an emergency.
You keep: the patched code, a record of what was fixed and why, and automated security checks running in your own repositories.
A four-week adoption pilot
The easiest way to start is one department, four weeks:
- Week 1: we read your existing policies, talk with the people who'll use the tools, note what data your staff work with, and draft your AI Use Standard.
- Week 2: a session for staff introduces the tools and the standard.
- Week 3: your team works under the standard. A second session, for developers or a follow-up for everyone depending on what week 2 showed.
- Week 4: we review what was produced against the standard and send a written report: what's working, what to change, and what to do next.
You keep: the standard, the quality procedures, both sessions' materials and the report.
What we don't do
We don't run your servers or networks. Systems administration stays with your staff. We work alongside them. We're not a security firm: we don't do penetration testing or run security monitoring.