1. Home
  2. Guides
  3. AI intake guardrails

Can AI handle law firm intake without giving legal advice?

Yes, as long as it is limited to collecting facts, running conflict checks, and scheduling. The guardrails are what make it safe: a narrow script, a clear disclaimer, attorney approval on anything substantive, and confidential handling from the first message.

By Xavier Iglesias · Updated October 31, 2026

What intake AI can safely do

  • Greet the person on behalf of your firm and say it is an assistant.
  • Ask your standard intake questions.
  • Check names against your conflicts list and flag possible matches.
  • Offer open consult times and book one.
  • Send your intake form and what to bring.

What it should never do

  • Answer questions about the law or someone's specific situation.
  • Predict outcomes or suggest what someone should do.
  • Quote fees beyond what you have approved.
  • Imply that an attorney client relationship exists.

Guardrails we set up

  • A short, approved script and a list of topics it must refuse.
  • A disclaimer in the first reply.
  • Escalation rules for urgent matters, like an upcoming court date.
  • Attorney approval before anything about a matter goes out.
  • Business accounts that keep data out of AI model training.

This guide is general information, not legal or ethics advice. ABA Formal Opinion 512 discusses lawyers' duties when using generative AI tools, and many state bars have published their own guidance.

Want this set up for you?

It starts with a free 30 minute call. You leave with at least one useful idea.

Start a conversation