Transform

Two out of three Australian small and mid-sized businesses now use AI in some form. Deloitte Access Economics surveyed more than 1,000 of them for Amazon in late 2025 and found that only 5 per cent are what they call fully enabled — AI embedded in the core processes, staff trained on it, data in one place. Everyone else is dipping a toe.

That gap is where the money is. Deloitte’s modelling put a number on it: moving from basic to intermediate use lifts profitability by about 45 per cent. Going from intermediate to fully enabled lifts it by 111 per cent. And if just one in ten Australian SMBs moved up one rung, the report estimated it would add $44 billion a year to the national economy.

The businesses that closed that gap did not do anything exotic. They picked specific workflows, wrote down the standard, trained the people, and kept iterating. What separates them from the businesses still experimenting is not the software. It is that they decided what the work should look like before they handed any of it over.

This section is the map: what an AI-native operation actually looks like from the inside, which roles to hand over first, and where to draw the line so nothing important leaves the building without a person signing off.

Write down one workflow you repeat every week, in the order it actually happens, and mark which steps need a judgement call.