How we build
- Spec: We write a specification, walk you through it and get your sign-off before any code is written.
- Rules: Our development standards are written down, and they apply to all of our development work.
- AI coding: AI coding agents, Claude Code and Codex, write the code under those written rules and review gates.
- Review: Every change is reviewed before it is merged.
- Checks: Tests run with the code, security tests included, and nothing ships until every check passes.
The software your business needs doesn't exist.
Sometimes nobody sells the software your business actually needs. We write the specification and design the software with you, and then we build it. When we're done, it runs in production on your own systems.
AI is writing your code, and nobody has set any rules.
When AI agents write code with no written rules and no checks, nobody can say with confidence what went into the product. We show you how to run them under a written, governed rulebook, with a review gate at each stage. Every change is then cleared before it goes in, and you keep a log that shows it.
Your code goes live without being tested.
If code goes straight to production with nothing tested, problems tend to be found by the people using it. We set up tests and security checks that run with the code, before any change is merged. Each release is cleared before it ships, and the check results are kept with the build.
Examples
Media Audit Group
live ·
Audit platform. Ian designed the audit model and built the platform its auditors use.
Meritum
in testing ·
Billing software for Alberta physicians. Ian Sharland built Meritum.
Clear Briefing
in build ·
A SigOct product. One attributed briefing from many overlapping sources.
LaunchLint
in build ·
A pre-launch security check for web and cloud projects.
Where to start
Tell us what the software needs to do, and we'll start with a fixed-fee written spec. Write to us at [email protected].
Questions
Who writes the code?
AI coding agents, Claude Code and Codex, write the code. They work under our written rules and industry-standard development frameworks, and every change is reviewed and tested before it is merged.
What is the first step?
The first step is a fixed-fee written spec. We walk you through it and you sign it off before any code is written, and then we build in milestones.
Can I buy Meritum, Clear Briefing or LaunchLint today?
Not yet. Meritum is in testing, and Clear Briefing and LaunchLint are still being built. The Media Audit Group platform is live, and Media Audit Group sells the audits.
Which platforms does the software run on?
It runs on Microsoft, Linux, AWS, Google Cloud and Azure.