TRAINING

Your team uses coding tools.
They need next-gen training.

Boundaries are unspoken. Review still assumes a human author. The only metric is a feeling in standup. We have put this practice across multiple teams, we have measured about 40% productivity gain on their own repos, and led a week long conference of training at Launch 2026.

WHAT GOES WRONG

Four common failures.

The tools write code fine.
The team process cannot keep up.

It writes more than anyone reads

Thousands of lines generated in a week. Review becomes a thumbs-up because reading them all properly would take longer than writing them did.

Review assumes an author

A reviewer reads a line and infers why somebody wrote it. Nobody wrote it. There needs to be a shift toward checks and balances and improved automated processes.

Nobody agreed the boundary

Old migrations, auth, payments, the deploy pipeline. Engineers made choices at a point in time, none of them wrote it down, or if they did, that documentation is severely outdated, and no two of them drew it in the same place.

The measurement is a feeling

The number of tickets closed went up. So did tickets reopened. Nobody is counting either one against what it was before the tooling arrived, so the argument in the room is between two anecdotes.

WHAT YOUR TEAM LEAVES WITH

Four artifacts.

All four survive us leaving. If your team cannot keep running this without a call back to us, the training did not work.

A working agreement

What the model may touch and what it may not, written down. Short enough that somebody remembers it at half past four on a Thursday, which is the only test a policy has to pass.

Review standards for generated code

What to check when there is no author to ask. Where tools are usually wrong, where they are confidently wrong, and which mistake costs more.

Guardrails in your pipeline

In CI, where they run whether or not anyone remembers them. A rule nobody can merge past is worth more than a rule everybody agreed to in a room.

A before and after on your repository

Measured against your code, over time. You get the grade whether or not it flatters the training.

SHAPE AND SCHEDULING

Some teams want this in a block, away from the work. Some want it beside the work, on the thing they are shipping that month. Your team's habits decide which, and so does the current workload. We work with you to determine what your people would benefit the most from.

When you engage with us we read your repository first. Teaching a team the general case and then discovering their real constraint on day three wastes the part of the training time which could have been targeted.

WHO IS TEACHING IT

Taught by someone who builds.

Rüadán Archer, Principal.
Twenty-eight years building and leading in software.

Team maturitymultiple teams have adopted this practice and have gained productivity
+40%productivity across those teams, measured on their own code
5 day conferenceteaching other software leaders at Launch 2026

LET'S TALK

Tell us what your team is already doing.

How many engineers, what they are shipping with, and what went wrong the last time. That is enough to say whether this is worth your budget.