The format, stepped through

The "AI Starter".

The idea

One AI tool, genuinely running. In about 30 minutes.

Most owners arrive having heard "you should be using AI", with no path from there to anything working. The AI Starter skips the possibilities and gets one useful tool running in your business this week. Here's the whole shape of it.

Think of AI as a brilliant, fast intern who is also a confident liar: great at the grunt work, as long as a person checks the result. That's the entire mindset.

Step 1 · Intake

What actually eats your time?

We start with your week, not the tools. A few honest questions:

  • What do you spend the most time on each week?
  • What would you happily hand to a fast intern, if you could trust the output?
  • Where do customers wait on you, after hours or when you're busy?

The answers point to the one or two places AI earns its keep for your business.

Step 2 · Pick one

One or two tools that fit your workflow. Not ten.

We choose what suits the answers above, then ignore the rest:

After-hours assistantA voice or chat line that answers enquiries and takes bookings while you're closed.
AI-assisted contentDraft emails, posts and replies fast, with a human edit pass that keeps your voice.
Document summarisingTurn long PDFs and notes into plain summaries (e.g. NotebookLM).
Light automationConnect the apps you already use so the busywork runs itself (n8n / Zapier).

Step 3 · First run

The thing you'll actually run this week.

You leave with a one-page plan: what to set up, and the first prompt or step to run. For an AI-content start, the first prompt looks like this:

You are helping [your business] reply to a customer enquiry.
Voice: warm, plain, Australian. Keep it short.
Here is the enquiry: [paste]
Draft a reply. Flag anything you're unsure about so I can check it.

Then you run it once, together, so it isn't theory. We also note the failure modes to watch for.

Step 4 · Check-back

Thirty days later: is it actually running?

A short follow-up answers three things:

  • Is the tool in genuine weekly use, or did it quietly stop?
  • Did it save the time we thought it would?
  • What's the next one small thing to add?

The goal was never a list of possibilities. It's one working result you can run on Monday, and a clear next step.

1 of 5