How I work · Stage 3 of 3
Build
Make it real, then check the data and refine. Continue, start, stop, pivot or kill, without drama.
These are stages, not phases. Discovery keeps running while you build, and the evidence gets collected the whole way through.
Build is where the choices become work. What do you need to make them real, what will we actually build, and how will we know it’s right? Then the part everyone skips. Did we reach the goal, and what does the data say, not the vibe?
I show rather than tell here. I get in the real work with your team, in your real tools, on your real tasks. People believe things they have watched get built. And AI does the mechanical lifting alongside us, so the human time goes where judgment lives.
The questions in this stage
Implementation
What do you need to make your choices real? Support, skills, structure to keep you on track. What will we build, and how will we know it’s right?
The artifact matters. A written operating model, a rebuilt workflow, a rewritten resume, a first version of the thing. Something that exists on Friday that didn’t on Monday, with a way to tell whether it worked.
Refine
Did we reach the goal? We check the data, not the vibe. Keep iterating. Continue, start, stop, pivot or kill, without drama.
Reinforcement is where most change dies. Small guardrails, a check-in, a number, someone who asks. Make the new way the easy way, then look honestly at what the numbers say and go again.
How it runs
For a business, I build alongside the team and leave the way of working behind when I go, with champions ready to run it. For a person, we build the plan into your actual week and keep a commitment log, because coaching works between sessions, not in them.
Then the loop. Fix, measure, go again. Big builds take time, but the small wins along the way compound, and the data will tell us when to continue, when to stop, and when to kill something with love.
Free tools that do this stage's work
The 6 Week AI Challenge, start here
The beginner path. Six weeks of small experiments in your own work, each week a different way of using AI. Showing it things instead of typing, handing it your files, saving a setup so you stop re-explaining, connecting it to where the work lives, and building a small tool. Pick the version for your role once you are in.
Use it if you keep meaning to get AI working for you and it keeps not happening.
Start the challengeBeyond chat, build your first AI agent
The advanced path. Six weeks from asking AI things to running a small team of agents that do the work while you check it. Delegation, giving it your context, connecting it to real systems, your first agent, putting it on a schedule, then a team with handoffs.
Use it if you already use AI every day and you have hit the ceiling of asking it questions.
Open the pathThe weekly check-in
Bring in your development plan and goals, then mark each week what actually moved, what only got touched, and what did not. It shows you the thing that has slipped three weeks running, which is the conversation worth having.
Use it between coaching sessions, or instead of one.
Open the check-inProduct interview questions, by level
Twenty-eight questions for product managers and product designers, split by level, from the person who used to ask them. Each one says what it tests, what a good answer sounds like, and the answer that means someone has read the books but not done the job.
Use it from either side of the table, whether you are hiring or preparing.
Open the questions