Is it safe to let an AI notetaker join your client meetings?

Only with everyone's consent, and in Australia that is a legal question rather than a courtesy. Recording laws vary by state and turn on consent, so the safe practice is to ask at the start of every meeting. Beyond consent, know where the recording and transcript are stored, who can reach them and how long they are kept.

The awkward feature of these tools is that they join quietly. A calendar invite carries the assistant along, it appears in the participant list with a name nobody recognises, and the meeting proceeds. Half the room may not notice; the client on the other end certainly will, usually about ten minutes in. The technology arrived faster than most businesses' habits around it, and the gap is closed by a sentence at the start of the call, not by a policy nobody reads.

Three questions settle the safety of any given tool, and they are worth answering before adoption rather than during a client complaint. Where does the audio go, your own Microsoft 365 tenant under your controls, or a third-party service under theirs? How long are recordings and transcripts retained, and who decides? And who can access them afterwards, since a transcript of a frank client conversation is exactly the sort of document that should not be sitting in a shared location by default. Tools that keep everything inside your tenant are easier to govern for the simple reason that your existing controls already apply.

The practical position we would take: a consent line said out loud at the start of every recorded meeting, a default of not recording anything sensitive, retention decided deliberately, and the tool chosen for where it stores data rather than for the quality of its summaries. Then treat the output as a draft, because a confidently wrong action item is worse than no notes at all. If you want meeting AI adopted with the consent and data questions answered first, call 1800 456 567.