// For.CTOs.whose.teams.code.with.AI
Your best engineer might be your biggest liability.
When AI-written code breaks in front of a customer, nobody asks which engineer wrote it. They ask who was leading engineering.
A personalized plan for your team, from a 2-minute survey.
ai-code-risk-guide · 12 questions · ~2 min
FreeQ1 / 12
If your board asked today how AI is used and controlled in engineering, you could:
Pick one to start. Eleven quick questions follow, and your guide shows what to tell your board. Your email is only asked at the end.
// The.real.risk
Your team isn't the problem. The missing system is.
Your engineers are doing what good engineers do: figuring it out on their own. But ten private ways of working with AI isn't adoption. It's exposure, and it sits on your desk.
- 01 The cautious engineer
- Barely uses AI. Leaves real gains on the table and becomes the bottleneck.
- 02 The fast engineer
- Trusts it too much. Ships more than the team can understand, review or safely change.
Opposite habits, same cause: nobody gave them a shared way of working.
// What.changes
Feel confident in every AI-written change your team ships.
Today everyone works with AI their own way, and you find out what went wrong after it shipped. We give your team shared methods and guardrails, so problems get caught before your customers see them.
- 01From guessing to confidence
- You know how your team works with AI, and you can trust what it hands back.
- 02From one star to a strong team
- The way your best engineer works becomes the way everyone works. Nobody is a single point of failure.
- 03From firefighting to reliable delivery
- Guardrails catch an agent saying "done" when it isn't, before a reviewer, your boss or a customer has to.
- 04From private tricks to a shared culture
- Your team learns from each other instead of reinventing the wheel, and the way you work keeps up when the tools change.
// The.course
Proven AI Patterns for Developers
The shared way of working, taught as a course. Five modules on prompting, context, evaluation and repository design, so every engineer works the way your best one does.
- Lessons
- 25
- Video
- 5h 53m
- Price
- $249 once
- M1
Mindset: the Compound Coder thesis
60m - M2
Accurate prompting
65m - M3
Context engineering
77m - M4
Evaluation and recovery
71m - M5
Agent-oriented repository organisation
80m
// Who.does.the.work
Troels Frimodt Rønnow
Principal engineer · 25+ years · ex-Microsoft
I ship production software with AI agents every day, in a large codebase behind more than forty automated checks designed specifically to make AI work reliable. I do every engagement myself, one team at a time.
Read my notes on troels.im// Free.guide · 2-minute.survey
Keep AI-written code from breaking in front of your customers.
Answer for your team as it works today. We use your answers to write your AI Code Risk Guide: where your team is exposed, how to close each gap for your stack and tools, and what to show your board.