Common questions Australian SMB owners ask us about using AI safely and effectively. Microsoft 365 Copilot, ChatGPT, hallucinations, data privacy, where to start, what to avoid. Direct answers first, then nuance — and a clear pointer to the AI Readiness Roadmap when there’s a structured solution. New FAQs added monthly.
Privacy & Data: “Will my business data be used to train Microsoft’s AI models?” (no — M365 Copilot stays in your tenant boundary), “What data should staff NEVER put into AI tools?” (anything Privacy-Act-covered, financials, health, NDA contracts — into public AI tools), “Does Copilot make us Privacy Act 1988 compliant?” (no — compliance is a posture, not a product). All three featured questions point back to Step 2 (Foundation) of our Roadmap, which configures the tenant boundary properly.
Adoption & Strategy: “Where should I start with AI in my business?” (start with the AI Readiness Assessment, not with buying licences), “How do I stop my staff using personal ChatGPT for work?” (an Acceptable Use Policy + tenant Copilot for the legitimate use cases — banning alone doesn’t work), “Will AI replace my staff?” (not in your business — but it will change what they do; train them or watch competitors lap you).
Quality & Hallucinations: “What about hallucinations — won’t AI just make things up?” (yes, sometimes — that’s why “verify don’t trust” is one of our three commitments). Copilot grounded against your SharePoint and OneDrive is materially less hallucination-prone than open ChatGPT, but it’s not zero. Treat AI output as a first draft, not a final answer — particularly for anything that goes to a customer or a regulator.
The full set of 7 Responsible AI FAQ pages is being progressively published. Each gets the same treatment: direct answer first, context, what Atria does. If you have a question we haven’t answered, email support@atriasystems.com.au. AI questions tend to compound — start with our free Copilot Readiness Assessment if you’d rather just have a conversation.
