Common Questions
Straight answers to what advisers actually ask — no jargon, no hype. If your question isn't here, a short conversation will sort it.
Yes. The obligations don't change because AI is involved — the advice, the best-interest judgment and the sign-off stay with the adviser. What AI does well is the work around the advice: surfacing information, drafting, chasing paperwork and preparing the groundwork a person then checks. Practices that do this safely decide up front which steps a human must approve, and keep a record of what the AI produced. Designed that way, AI sits comfortably inside your licence rather than testing its edges.
A product is built to be sold to thousands of practices, so it solves the average problem and assumes you'll do the work of fitting it in. Consulting starts from your practice instead — what you run, where the hours actually go, and which parts of the week are worth changing. Sometimes the answer is a tool you already own, used properly; sometimes it's connecting two systems that don't talk to each other; occasionally it's something new. The recommendation is shaped by your practice, not by what a vendor happens to sell.
Somewhere small and unglamorous. The best first project is a task that happens constantly, follows the same steps each time, and quietly eats hours no one enjoys — the paperwork around consents, or the groundwork before a client review. Starting there keeps the risk low and the disruption minimal, and it proves the idea on something concrete before you commit to anything bigger. Ambition can come later; the first move should be safe and obvious.
It can be — but safe is a design decision, not a default. It comes down to where data is processed and stored, what a given tool keeps or learns from, and drawing a firm line between practice systems and the free consumer AI tools that may train on whatever you type into them. A large part of the work is choosing the right tools and shaping the workflow so client information is handled the way your obligations require — deliberately, rather than hopefully.
Usually not, and the distinction matters more in a licensed practice than almost anywhere else. An automation follows a track you laid out in advance, so it does the same thing every run and can be walked through step by step. An agent works out its own steps as it goes. The question that separates them is simply who decides what happens next. Most of what gets pitched as an agent turns out to be a fixed sequence with one small AI step in the middle, which is cheaper, steadier and far easier to explain to your licensee. The full guide walks through the question, three worked examples and the two things that trip people up.
Not at all. You understand your practice — that's the part that can't be outsourced. The technical translation is the job of the person you bring in: turning "AI" into specific, plain-English changes to how your week runs, and handling the build so you don't have to. If you're not sure where you stand, the free AI-readiness check is a low-pressure way to get a read, and a first conversation goes from there.
Still Weighing It Up?
Thirty minutes, no pitch deck, no obligation — just a straight read on where AI fits in your practice.
Prefer email? david@baxadvisory.com.au