Software development for Australian businesses

Operational systems that have to keep working where connectivity does not, and reporting that has to survive an audit years later. The recurring engineering problem is reconciliation: data captured on a site with intermittent connectivity has to merge cleanly once a device is back in range, without a person deciding which version won. Energy also sits inside both the critical infrastructure regime and the Consumer Data Right, so the compliance surface is unusually wide for what often begins as a straightforward internal tool.

Citizen-facing services and the systems behind them, where the procurement process asks harder questions than the brief does. Accessibility is generally assessed rather than assumed, security expectations are stated up front rather than negotiated, and the record of who accessed what tends to matter as much as the feature itself. Building to those expectations from the first sprint is far cheaper than retrofitting them in response to an assessment.

Member portals, reporting layers and the reconciliation engines underneath them. Correctness under partial failure decides these builds, because the failure modes are quiet: a balance that is right on the screen and wrong in the ledger is worse than an outage, and it is usually discovered by a member rather than by monitoring. Financial services also sits inside the critical infrastructure regime, which turns incident detection into a product requirement rather than an operations concern.

Clinical and patient-facing systems where the boundary matters more than the feature set. Which services hold health information, which merely reference it, and how the audit trail stays intact across that line are decisions that set the compliance surface for everything built afterwards. Health care and medical is also one of the eleven critical infrastructure sectors, so the twelve-hour reporting expectation applies to organisations that do not usually think of themselves as infrastructure.