Insights
Perspectives on the technology decisions that materially affect enterprise outcomes — written by the people doing the work.
The gap between an impressive prototype and a reliable AI system is engineering discipline. Evaluation harnesses, retrieval quality and failure modes decide whether a model survives contact with real users.
Most automation projects degrade quickly because they were never designed to be maintained. Ownership, observability and versioning make the difference between compounding value and compounding debt.
Programmes that treat transformation as a technology procurement fail. Programmes that treat it as an operating-model change, executed through software, succeed. What separates the two.
The technical decisions matter less than the operating rhythm of the team. Pipelines, contracts, testing and cadence — the unglamorous practices that reliably predict delivery.
A working guide to the shifts durable enough to shape architecture decisions today — and the ones safe to ignore. Platforms, AI as substrate, data consolidation and developer experience.
Replatforming rarely fails on technology. It fails on sequencing, data migration and the invisible workflows that keep the current system running. A pattern for staged modernisation.
A decade in, most enterprises are neither saving money nor moving faster in cloud. The gap between the promise and the outcome is an operating-model problem, not a technology one.
The strongest security postures come from teams that treat security as a design constraint, not a compliance exercise. Identity, secrets, supply chain and detection — the practices that matter.
Practical writing on applied AI, enterprise modernisation, cloud economics and delivery — from the engineers building production systems for regulated Australian organisations.
We work with a small number of clients each year. Tell us what you're building.