The wrong opening question is usually about the tool.
Many leadership conversations begin with a product: should we roll out a copilot, build an agent or automate a workflow? Those questions matter later, but they are a weak place to begin. They encourage a technology decision before the business has agreed what should improve.
A stronger opening question is: where is the organisation losing time, quality, confidence or commercial momentum—and what would meaningfully change if that constraint eased? That keeps the work anchored to a real operating outcome.
Choose a problem that is important enough to own.
The best early AI opportunities are not necessarily the most technically impressive. They are specific, visible and owned by someone who can make decisions about the process around them. A pilot without a business owner quickly becomes a demonstration rather than a capability.
For a Perth or WA business, the right starting point may sit in customer response, proposal development, knowledge retrieval, reporting, project delivery or repetitive operational coordination. The point is not to make every task ‘AI-enabled’; it is to improve a result people already care about.
- What business result should improve?
- Who owns the process and can change the way work is done?
- What information can be used safely and with appropriate controls?
- How will the team know the work is genuinely better?
Treat adoption as part of the strategy.
A useful use case still fails if people do not understand when to use it, what good looks like or where responsibility sits. The strategy therefore needs a simple operating model: accountable owner, intended users, guardrails, support, measures and a decision point on whether to scale.
This is where human-led, AI-assisted work matters. AI can accelerate analysis, drafting and workflow steps. People still set the outcome, judge quality, manage risk and decide when a process is ready to change.
A sensible first move.
Start with a short opportunity assessment rather than a large programme. Bring together the business owner, people who do the work, the relevant technology or risk perspective and one decision-maker. Identify a small number of opportunities, compare them against value and feasibility, then choose the first move deliberately.
The output should be practical: a prioritised use-case list, a clear first experiment, named ownership, basic safeguards and a way to measure whether the change is worth continuing.
The goal is not to appear advanced. It is to make a better business decision about where AI can create useful, durable value.