An AI strategy in dollars, not a deck that dies in a drawer.

What should an AI strategy actually contain?

A ranked list of the processes worth automating, each with a dollar estimate, a named owner and a 90-day path to live. It should fit in one sitting, say plainly what to do first, and say just as plainly what is not worth doing yet. If you cannot act on it the week you receive it, it is not a strategy.

The AI strategy most Australian businesses are sold is a transformation document: market context, maturity models, a vision statement, a roadmap with horizons. It reads well in the boardroom and then it dies in a drawer, because nobody can point to the first thing to build, what it is worth or who owns it. We write ours the way an operations manager would want it written: process, number, owner, sequence. The shape holds at 50 staff or 500, down the road from our Ardross office or three time zones east.

That shape is deliberate. A mid-size business does not need a transformation program. It needs its first three automations to pay for themselves, because once quoting, purchase order chasing or the monthly reporting pack is demonstrably faster and cheaper, the argument for the next three is already won. Good strategy is mostly the discipline of choosing what not to do yet.

How do you put a dollar figure on AI?

With arithmetic, not industry benchmarks. How long the job takes, how often it happens, what that time costs at real wages, less what the automation would cost to build and run. We use your numbers, we round down, and we label every figure an estimate, because that is what it is until the build proves it.

Take quoting in a hire or services business. If a quote takes two hours, the office produces several hundred a year, and half of that time is looking things up and formatting rather than judgement, the value of automating it is a multiplication you can check yourself. We run the same sums on purchase order chasing, supplier prequalification packs, follow-up emails and the reporting that swallows the start of every month. Some jobs come out of the exercise looking underwhelming, and they go on the not-yet list with the reason written next to them. That list is half the value of the plan.

We are equally plain about what the arithmetic cannot promise. Freed hours only become dollars if you redeploy them, revenue effects are real but slower to prove, and a process that changes every month is a poor candidate no matter how the sums look. An estimate dressed up as a guarantee is how AI strategies lose the room, so we never dress them up.

What is the difference between AI strategy, an AI audit and AI consulting?

Three depths of the same thinking. The AI audit is the fixed half-day entry point that finds your first handful of opportunities and comes back with a costed roadmap within days. Strategy is the full plan: the whole operation ranked, costed and sequenced. Consulting is the ongoing thinking once the plan meets reality.

Most clients start with the audit because it is quick, fixed in scope and forces an early decision instead of a long engagement. When the audit shows more than a handful of processes worth doing, strategy is the natural next step: the whole business looked at once, so effort goes where the return is rather than where the noise is. And once things are being built, consulting is the standing conversation about what to do next, what to stop, and what the last build taught us. Everything sits on our services page, but those three are the thinking layer, and the same people do all of it.

What if the right strategy is to wait?

Then the plan says wait, and says what to fix first. If your numbers live in heads and inconsistent spreadsheets, if you are mid-way through changing core systems, or if no process clears its own cost, automating now just makes the mess move faster. An honest strategy includes the no.

Waiting is not the same as doing nothing. A wait verdict usually comes with small, cheap homework: fix this register, standardise that template, get quoting out of one person's head, then revisit in six months. Often the right first move is people rather than software, a session that equips your board and executives to judge AI proposals on their merits, which is exactly what our executive and board briefings are for. We are independent, nobody pays us to push a product, so we have no reason to tell you to buy anything this quarter.

And when the answer is go, the plans get built. Every build on our success stories page started life as exactly this thinking: a ranked look at where the hours were going in a real WA business, then the first automation, then the next. The names and photos are there because a strategy is only as good as what it turned into.