Service

Our share of the commits falls.
Yours rises, week by week.

A one-week, on-site workshop that teaches your developers to build production software with AI agents. It ends with a functional agentic setup, not slides. An optional eight-week guidance phase then follows, inside your own codebase. Your developers lead that work from day one.

See the proof
Our share falls, yours rises

The workshop sets up the guardrails. Lint rules, complexity limits, and size caps go into the repository.

1 week
on site: prep, lessons, and hands-on work
2,500+
commits in the six weeks after the workshop
~75%
of that new codebase written by the client's own team
120+
versioned database migrations in the new codebase

01 - The situation

Everyone uses a chat window. Few trust it in production.

Your developers already try an AI assistant on their own. That does not yet make the codebase safer or faster.

Where agentic coding stalls

  • A developer copies code from a chat window, with no guardrails and no review rhythm.
  • The code looks right, but nobody set a lint rule, a complexity limit, or a size cap.
  • One senior developer knows the method. The rest of the team copies it by guesswork.
  • An external team builds the system fast. The pace drops the day that team leaves.

What we do instead

  • We teach your own developers the method, on your machines, inside your own stack.
  • We commit lint rules, complexity limits, and size caps to the repository, on day one.
  • Every developer in the room learns the same method, not only the senior engineer.
  • Your developers lead the build. Our share of the commits falls week over week, by design.

02 - What you get

Three things the workshop leaves behind

Not a slide deck and not a proposal. A method that works, in your own repository, that your team keeps after we leave.

01

A functional agentic setup, on every machine

Every developer leaves the week with a functional agentic harness, tuned to your own stack.

What it is

We install and configure the harness on each machine during the hands-on days. We tune every setting to your languages, your frameworks, and your existing repository.

Why it matters

The guidance phase then starts with a tool that your team already trusts. Nobody installs anything on day one of stage two.

02

Guardrails and a review rhythm, in your repository

A guardrails configuration and a practised method to review agent-written code, both committed to the repository.

What it is

We commit lint rules, complexity limits, and file and function size caps to the repository. We then run the review rhythm live, on your own pull requests.

Why it matters

The rules survive the workshop. Every commit after the week meets the same bar, with no reliance on memory.

03

A guided build, inside your own codebase

Eight weeks of pair sessions, code reviews, and architecture decisions, with your developers in the lead.

What it is

We work next to your team, inside the codebase that it now owns. We pair on real problems. We decide the architecture together.

Why it matters

Our share of the commits falls week over week. Your team keeps the pace after the engagement ends.

03 - How it runs

Six weeks, and the balance flips

The workshop sets up the method in week one. Your developers carry more of the load every week after it.

Our share falls, yours rises

The workshop sets up the guardrails. Lint rules, complexity limits, and size caps go into the repository.

Week 1

The workshop, on site

The week holds a prep day, two lesson days, and two hands-on days. Every developer leaves it with a functional setup and the guardrails already in the repository.

Weeks 2-3

First agentic slices

Your team builds against the specifications from the hands-on days. The consultant's share of the commits already begins to fall.

Weeks 4-5

The pace climbs

Guardrails and the review rhythm hold under real delivery pressure. The team's own commit count climbs week by week.

Week 6

The team leads

The client's own developers now write most of the commits. In one engagement, that share reached roughly three quarters within six weeks.

04 - Proof

Where we did this work

One engagement, told in full: the workshop, the guardrails in the repository, and six weeks of commits after it.

A team learns to build with agents: one workshop week, then a production port at startup speed

Commodity trading, circular economy

A team learns to build with agents: one workshop week, then a production port at startup speed

One week of on-site agentic development training, then eight weeks of guidance. A scale-up's own developers ported their internal operations platform from an AI app builder to a production TypeScript stack. The port took more than 2,500 commits in six weeks, and the team itself wrote roughly three quarters of them.

Read the case study
All case studies

05 - Questions

Common questions

What does agentic coding training actually teach?
Agentic coding training teaches your developers to direct an AI agent through a full task, not only to chat with a model. The workshop covers the agentic harness, the guardrails that a repository needs, and a review method for agent-written code. It ends with a functional setup on every machine, and a guardrails configuration already in the repository. In one engagement, the client's own developers wrote roughly three quarters of a new production codebase within six weeks of the workshop.
Is this an AI coding workshop for developers, or a workshop for managers?
The workshop is for developers. Every lesson and every hands-on exercise runs inside your own stack, not a generic slide deck. The two lesson days cover 14 topics: the basics of an LLM, safe work with guardrails, the daily toolkit, and the disciplines that compound. The two hands-on days apply all of it to your own system, with a real specification as the result.
Do you offer AI-assisted development consulting after the workshop?
Yes, as an optional second stage. We pair with your developers inside their own codebase, review pull requests with them, and decide the architecture together. Your developers lead the work from day one of that stage. Our share of the commits falls week over week, by design. Your team's pace rises in its place.
What AI coding guardrails go into the repository during the workshop?
We commit lint rules, complexity limits, and file and function size caps to the repository. We add agent instruction files at the repository root, and in every major area. Every agent session then starts with your conventions already in context. We also set database changes to run only through versioned migrations. All of it lands inside your repository, not inside a slide deck.
Is this developer upskilling on AI agents, or do you write the code for us?
It is developer upskilling. Your own developers write the code, in the workshop and in the guidance phase that follows it. In one engagement, the client's developers wrote roughly three quarters of a 2,500-commit codebase within six weeks of the workshop. We teach the method. We then guide the first build.
How does this support engineering team AI enablement, not just one developer?
The workshop trains every developer on the team, not one champion. Guardrails and conventions go into the repository, so the practice does not depend on memory. A documented multi-agent worktree workflow lets each developer run several agent sessions in parallel. A team that finishes the week keeps its pace after the engagement ends.

06 - Related

Other services

Most engagements combine two of these. Start where the risk is highest.

MVP development sprintA fixed-scope sprint of 4 to 6 weeks. It ends with a deployed product, the codebase, and a roadmap.Custom-built AI agentsAgents that take real actions in your product and your operations, under permission controls you select.Custom software and modernizationOne system that replaces the spreadsheets, or a ground-up rebuild of a product you still run.Long-term product partnershipA senior team that builds, operates, and extends your product every month, at fractional cost.Proof of concept and pilotsA functional prototype in 3 to 6 weeks that answers one question: scale it, or stop it.AI opportunity workshop and roadmapA workshop with your process owners that shows where AI pays off, and in which order.Technical due diligenceA review of the code, the architecture, and the team before you invest, delivered as a written report.
This service, written for one audienceFor startups

Ready to teach your developers the method?

Book a workshop scoping call. We tailor the week to your stack. You leave with a concrete plan and a concrete price. Stage two stays your choice, on evidence, after the workshop ends.

One week on siteConfidentialReply within one business day