One awkward disclosure at the top of a call can cost you a client. A bot sitting in the corner of a video meeting, a transcript saved to a vendor's server, a summary that reads back a sentence the client never said — any of these can turn a routine intake into a bar complaint, a privilege fight, or a lost engagement. Small firms absorb that risk most directly, because the partner who signs the retainer is usually the same person clicking 'record.'
The useful question is no longer whether to bring AI into client work. It's where in a conversation the tools genuinely help, where they should only clean up afterward, and where they don't belong in the room at all.
Before the Call, AI Earns Its Keep Quietly
The hour before a client call is where a small firm has the most to gain and the least to lose. Nothing has been said yet. No privileged conversation is being captured. The work is preparation, and preparation is the kind of task where a well-configured tool saves real time without dragging ethics into the picture.
That prep window is where growing small firms are pulling ahead. Clio's latest trends report found that growing firms use time-saving automations — booking, drafting, summarization, task prioritization — roughly twice as often as stable firms and nearly three times as often as shrinking ones. The gain isn't magic. A partner walks in already briefed.
During the Call, Assume the Client Can Hear the Robot
The live conversation is where the risk profile changes. A recording or transcription tool sitting in an attorney-client call is not the same as one sitting in a sales demo. Consent laws vary by jurisdiction, privilege can be brittle when a third-party vendor sees the audio, and clients notice when a bot joins.
The New York City Bar's Professional Ethics Committee took this up directly, and its formal opinion on AI recording and transcription of client conversations lands on a straightforward rule: get informed client consent, weigh where the audio and the resulting text will live, and verify what the tool produces before you rely on it. That is the floor, not the ceiling.
There's also a softer cost that never shows up in an ethics opinion. Clients speak differently when they know they're being recorded. The nervous business owner walks back the sentence that would have told you what happened. The spouse in a family matter stops mid-sentence.
That candor is often the whole point of the meeting. If the tool is visibly in the room, the meeting you booked isn't the meeting you get.
After the Call, Let AI Do the Cleanup
The post-call window is where a lot of the honest productivity lives. A dictated summary becomes a structured file note, a follow-up email gets drafted from bullet points, and the task list routes itself to the right paralegal. Deadlines land on the calendar. None of it requires the AI to have been in the room, and all of it makes the practice more responsive — which is what clients actually remember.
That responsiveness matters more than most firms admit. Longstanding professional-conduct guidance expects a lawyer to meet reasonable client expectations for information, consistent with the client's best interests. A same-day follow-up note, drafted from the attorney's own dictated debrief and reviewed before it goes out, is a direct answer to that duty.
A serious caveat sits under all of this: the attorney owns the output. Every summary needs a read. Every drafted email needs a signature that means something. If a tool misquotes what the client said and it lands in a file note, the file note is wrong, and it's wrong under the lawyer's name.
Draw the Line in Writing, Then Tell the Client Where It Is
The firms handling this well are boring about it. They decide, in writing, which stages of a matter can involve AI, which cannot, and what the client will be told. That policy lives in the engagement letter, gets reintroduced at intake, and gets honored on every call.
That pragmatic posture is exactly what a good field guide describes when it walks through using using AI in a small practice without losing the human touch in a small practice without losing the human touch.
The client call is still where the practice happens. AI can shorten the walk to it and clean up after it. Inside the call, the lawyer still has to show up as a person — and be willing to close the laptop when the moment calls for it.
