How we work

Start with the problem

We work out what needs building and why it's worth building before writing the code. More complex work is split into phases. You see something working at each stage and review changes before they go live.

Under the bonnet

Your application isn't a testing ground

We develop locally, provide a test version for you to review, and release approved changes to the live application.

Every change has a date and an author

The code lives in a Git version-controlled repository on GitHub. We can see who changed what and return to an earlier version when needed.

Hosting and ongoing care

We maintain servers across multiple datacenters in Australia, and deploy to AWS, Google Cloud or Azure when a project calls for it. We add a content delivery network when needed. We take care of hosting, updates and backups, and know where the application runs.

The tools

We invest an inordinate amount of time making sure we pick the best and most reliable tools for each function.

Today that means more than fifty specialised platforms, from graph databases like Neo4j and AI frameworks like LangChain to Databricks for data, Vercel for hosting and ElevenLabs for voice.

The list changes as better options appear. Experience helps us decide what fits the job and what will be practical to maintain.

Speed

Speed matters. Smart companies are already integrating AI at full speed. We know, some are our clients.