Walk into ten boardrooms in Sydney this quarter and you'll hear the same sentence: "We're getting our AI strategy together." Press for detail, and what you usually find is a Microsoft Copilot trial, a Slack channel called #ai-experiments, and a director who watched a Dario Amodei interview on the weekend.
None of that is readiness. Readiness is operational.
Signal 1: Someone owns it
Not a working group. Not a steering committee. One named human, with budget authority and a calendar that reflects the priority. If you can't tell us who owns AI in your business in under ten seconds, you're not ready, you're interested.
Signal 2: Data is somewhere it can be used
The fastest way to kill an AI program is to discover, six weeks in, that the data you need lives in 14 spreadsheets, a legacy CRM, and a fax-server-shaped object in the corner of the finance department. Readiness includes having data that's accessible, classified, and governed enough that a model can touch it without legal sending you a memo.
Signal 3: You have a use case smaller than a moonshot
The single best predictor of AI program survival is whether the first use case can ship in 90 days. "Transform customer service across the enterprise" cannot. "Auto-triage inbound support emails in the consumer division" can. Pick the second one.
Signal 4: Adoption is in the plan, not in the postscript
If your AI program plan mentions training, change management, and feedback loops on page one, you're ready. If those appear in week 11 of the Gantt chart as "rollout", you're going to ship a system nobody uses.
Australian businesses are good at execution when they pick the right battles. The AI battle is winnable, but only for the ones who stop confusing curiosity with capability.