How an engagement runs, step by step.
This page is the procedure, from the first call to handover. It covers who you talk to, how the work is scoped and staffed, how often you see it, and what you hold at the end.
Before work starts
- 01
The first call
Leo Mata, our founder and CEO, usually takes the first call. You describe the problem, the systems involved and what a good result would look like. Bring your technical people if you can. - 02
Discovery and proposal
Discovery is structured. We record and analyse the calls, then write a proposal. It sets out what we learned, the approach and architecture, the scope as numbered workstreams, a phased plan, the team by role, and what is included and what is not. Estimates are in hours and given as ranges, and most proposals offer two or three named options.
For a large build we propose a short paid scoping phase first. A senior team investigates your systems and produces a validated plan and an estimate. No code is written in that phase.
- 03
Contract and team
We sign one master services agreement, then a statement of work for each piece of work. The proposal ends with a list of what we need from you to start, which is access, credentials, decisions and time with the right people. When a scoping phase came first, the team that did it carries on into the build.
While we build
- 01
A live version early
New builds start on Boilerworks, our open source templates, with authentication, permissions, CI/CD and infrastructure already wired. The first goal is the shortest path to a live URL. Everything after that is iteration you can see. - 02
Sprints and demos
We work in sprints of one to two weeks, and you review and approve what goes into each sprint at the planning session. We demo working software every two weeks or every month. Between demos you have live deployments and shared documents to look at. - 03
When the scope changes
On an embedded engagement you set the priorities again at each sprint planning session. On fixed-price work, input that changes the scope is re-scoped. A change that alters the cost or the timeline of the agreed work goes through a written change order.
How the code gets written
- 01
Plans before code
Work is broken down into ordered tasks, each small enough to be one reviewable pull request. The plan for a task is written on its issue before any code, and the issue is closed with a summary of what was done. - 02
Review and quality gates
Every change is reviewed. Lint, tests, build and a dependency audit run in CI, and a failure blocks the merge. Tests run against a real database and cover the error and permission-denied cases as well as the expected one. - 03
AI agents and senior engineers
AI agents do much of the repetitive build work under the same rules. An agent reads the issue, posts a plan, works on a branch, writes tests and opens a pull request that has to pass review. Senior engineers oversee the work and make the architecture, integration, security and operations decisions. Navegador and Scuttlebot, two of the tools we built for this, are open source.
Handover and after
- 01
Documentation and handover
Documentation is written alongside the code and kept in the repository. When your team will run the system, handover is part of the scope. That means runbooks, architecture documentation and working sessions with the people who will own it. - 02
What you hold at the end
You hold the code, delivered to your source control, and its documentation. Under our master services agreement you own the work product we create for you once it is paid for in full. We keep our own pre-existing tools and methods and license to you any that are delivered as part of the work. Boilerworks, the foundation our new builds start on, is open source under the MIT licence, so your engineers can read it before you sign. - 03
Operations after launch
Support, on-call and site reliability work after launch are engagements in their own right. Team Spartan, our DevOps, SRE and security group, supports production systems on a monthly retainer and responds to incidents. Its prices are public.
Why we work this way
This page describes what we do. The reasoning is on two other pages. Our Approach explains how we build around AI agents, and Outcome Engineering explains why every engagement starts from the business result it has to produce.
Start with a call
The first call is usually with Leo Mata, our founder and CEO. If you would sooner write, use the contact form.


