A prospective client calls a law firm after hours asking a specific question: "Do I have a case?" Whoever — or whatever — answers that call is now one sentence away from a serious problem, because answering that question is legal advice, and legal advice requires a license.

This is exactly why most law firms are cautious about automating intake at all. The caution is correct. The conclusion that automation isn't possible is not.

The real risk isn't automation — it's an AI that improvises

Unauthorized practice of law doesn't happen because a firm used software. It happens when something — a person or a system — gives legal advice, predicts an outcome, or characterizes someone's situation without a license to do so. A generic chatbot vendor with no legal context will eventually improvise an answer to sound helpful. That's the actual liability, not the automation itself.

What's safe to automate in a law firm's intake

The line is narrower than most firms assume, but it's real and it's workable:

  • What it handles: after-hours call answering, scheduling consultations, collecting basic intake facts (name, matter type, contact info, referral source), missed-call text-back, appointment reminders and confirmations.
  • What it hands off immediately: any question about merits, likely outcomes, deadlines, or "what should I do" — routed straight to an attorney or paralegal, with zero attempt to answer.
The difference between a compliant intake system and a liability isn't how smart the AI sounds — it's whether it was built by someone who understands exactly where "collecting information" ends and "giving advice" begins.

Why generic AI vendors get this wrong

Most automation vendors optimize for the AI sounding capable, which is precisely backwards for a law firm. The right build does the opposite: it's deliberately conservative, defaults to "let me connect you with someone" the moment a question moves past pure intake, and never guesses. That boundary gets agreed with the firm in writing before anything goes live — not left to the AI to figure out in the moment.

What this looks like in practice

A firm running this setup typically sees after-hours consultation requests get captured and scheduled instead of lost to voicemail, missed calls during court or client meetings get an instant text instead of silence, and intake information arrives complete and organized before the first human conversation even starts.

None of it replaces the judgment a law license exists to provide. It just makes sure that judgment gets applied to every lead, instead of the ones who happened to call back a second time.