dialessa Have it call me

Guide

AI receptionist for Australian service businesses: answer every call, stay compliant

An AI receptionist answers your inbound calls when you can't. It greets the caller, works out what they need, books them in or takes the details, and passes anything complex to a person. For a service business that misses calls while on the tools, with a client or after hours, that can mean fewer enquiries going to voicemail and then to a competitor.

The technology is only part of the job. In Australia you also need to think about telling callers they're talking to AI, saying when a call is recorded, and giving a privacy notice when the call collects personal information. This guide covers both the practical setup and those obligations.

What an AI receptionist does on a call

A good AI receptionist follows the same steps a skilled front-desk person would. Plan each step on purpose rather than leaving it to the defaults.

Greeting

The opening few seconds set the tone. The greeting should name your business, say the caller is speaking with an AI assistant, and mention recording if you record. Keep it short and natural. Callers who rang to fix a leaking tap or book an appointment want to get to the point, not sit through a long preamble.

Qualifying

Next, the agent works out what the caller wants and whether you can help. For a service business that usually means:

  • What service they need
  • Where they are, if you only cover certain areas
  • How urgent it is
  • Any details that affect price or scheduling
  • Their name and the best way to reach them

Write these questions down before you configure anything. The aim is to collect what your team needs to act, and nothing more. Every extra question adds friction, and every extra piece of personal information is something you have to protect.

Booking

If the agent can see your calendar or job management system, it can offer real time slots and confirm a booking during the call. If it can't, it should take a request and tell the caller clearly what happens next and when. Don't let it promise a time it can't guarantee. A confirmed booking that turns out to be wrong does more damage than an honest "someone will call you back this morning".

Handing over to a person

Decide in advance when the AI should step aside. Common triggers include:

  • An emergency or safety issue
  • A complaint or an upset caller
  • A question about price, warranty or contract terms that needs judgement
  • The caller asking for a human
  • The agent not understanding after a reasonable attempt

Handover can be a live transfer during business hours, or an urgent message to whoever is on call after hours. Whatever you choose, the caller should never feel stuck in a loop with a machine.

Telling callers they're speaking with AI

Modern voice agents can sound very human, which is why disclosure matters. The National AI Centre's Guidance for AI adoption says people should know when they are interacting with AI. Its example is a customer getting services from a chatbot, who should know it is not a human specialist. The same guidance recommends making AI disclosure standard practice, especially where AI can be mistaken for a human being or where it makes or influences decisions.

The privacy regulator takes a similar view. The OAIC's AI guidance says that when an organisation collects personal information through a public-facing AI system, it should make sure people are aware they are interacting with an AI system rather than a human.

In practice, this is easy. Put it in the greeting: "Hi, you've reached [business name]. I'm the AI assistant and I can help book a job or take a message." Callers are usually fine with this when the agent is quick and useful. What annoys them is finding out halfway through. If a caller asks "Am I talking to a real person?", the agent should always answer honestly.

Announcing that the call is recorded

Most AI receptionists record or transcribe calls. That's how you review quality, check what was agreed and fix mistakes. Recording laws in Australia differ between states and territories, and they don't all work the same way:

Your callers may ring from anywhere in the country, so the simplest approach is to announce recording at the start of every call. A short line such as "This call is recorded for quality and booking purposes" gives callers the chance to object or hang up. It also supports an argument that people who stay on the line have consented. Whether an AI system counts as "a party" to a conversation, and how these laws apply to your setup, is a question for a lawyer. Get advice if recordings will be shared outside your business or used for anything beyond running the service.

The privacy notice when a call collects personal information

Almost every useful receptionist call collects personal information: a name, a phone number, an address, sometimes details about health or property. Whether the Privacy Act applies to you depends on your business.

The OAIC says most small businesses, meaning those with an annual turnover of $3 million or less, are not covered by the Privacy Act 1988, but some are covered regardless of turnover, including health service providers and businesses trading in personal information. Allied health clinics, dental practices and similar service businesses should assume they're covered.

If you are covered, Australian Privacy Principle 5 requires you to take reasonable steps, at or before collecting personal information or as soon as practicable afterwards, to tell the person matters such as your identity and contact details and why you're collecting it. On a phone call, that could be a brief spoken notice plus a link to your full privacy policy in the confirmation text or email.

The OAIC also recommends businesses update their privacy policies and notifications with clear information about their use of AI, including clearly identifying public-facing AI tools to customers. Your privacy policy should say you use an AI phone assistant, what it collects and who processes the data.

Look ahead too. From 10 December 2026, organisations covered by the Privacy Act that use personal information in automated decision-making that could affect people's rights or interests must explain in their privacy policies the kinds of personal information used and the kinds of decisions made. If your AI receptionist decides who gets an urgent slot or who gets turned away, check with a lawyer whether that applies. The stakes have also risen: the Privacy and Other Legislation Amendment Bill 2024 introduced a statutory tort for serious invasions of privacy and gave the OAIC new tiers of civil penalties and the power to issue infringement notices.

Even if you're a small business outside the Act, following these practices is good for trust and keeps you ready if your circumstances change.

A note on follow-up marketing

If you plan to use details collected by the receptionist for later promotional calls or messages, other rules may apply. The OAIC says APP 7 on direct marketing generally applies to direct marketing calls to numbers not on the Do Not Call Register, but not to communications covered by the Do Not Call Register Act or the Spam Act. Where APP 7 applies, you must give people a simple way to opt out and stop using their information for direct marketing within a reasonable period if asked. Check with the ACMA or a lawyer before turning inbound callers into a marketing list.

Setting up an AI receptionist: practical steps

  • Map your most common call types and what a good outcome looks like for each.
  • Write the greeting, including AI disclosure and the recording announcement.
  • List the qualifying questions and cut any you don't need.
  • Connect the agent to your calendar or job system, or define a clear callback promise.
  • Set handover rules and who receives urgent after-hours alerts.
  • Draft a short spoken privacy notice and update your written privacy policy.
  • Decide how long recordings and transcripts are kept, and who can access them.
  • Listen to real calls each week at the start and refine the script.

Questions to ask a provider

  • Where is call data stored, and who can access it?
  • Can I edit the greeting, disclosures and questions myself?
  • How does live transfer work, and what happens if nobody picks up?
  • Can I delete recordings and transcripts on request?
  • Does it integrate with the booking tools I already use?
  • How does it handle callers who are upset, confused or asking for a human?

Putting it together

An AI receptionist works best when it's honest from the first sentence: it says it's an AI, says the call is recorded, collects only what it needs and hands over smoothly when a person is better placed to help. Done that way, you can answer every call around the clock without losing the trust that keeps a service business going.

If you'd like to see how this works in practice, Dialessa is an AI voice sales agent that answers inbound calls, calls new leads within seconds and handles follow-up, so fewer enquiries slip through.

This article is general information, not legal advice. Rules change; check the current requirements with the ACMA and the OAIC, or a lawyer, before you rely on them.

← All posts