Are AI intake chatbots ethical for law firms?

What must an intake bot never do?

Four things: apply law to the visitor’s facts (that is legal advice and potentially unauthorized practice), promise outcomes, contradict advertising rules with promises of results or comparative claims, and misrepresent itself as human or as an attorney. Every one of these is a system-prompt fence, testable before launch.

What can it safely do?

Everything administrative and educational: explain the firm’s process and fees, collect matter type, timeline, and contact details, answer general questions about what the firm handles, run conflict-screening questions, and schedule consultations. That covers the entire job of intake without crossing into advice.

What does a compliant deployment look like?

Clear disclosure on the widget (“AI assistant: not a lawyer, no legal advice”), a fenced system prompt with refusal behaviors, human review of conversation logs, secure data handling for intake details, and an escalation path to a person. Deployed this way, the bot is an ethics improvement over voicemail: it never forgets a disclaimer.

Related next steps

• AI Policy Generator

  Draft a written AI use policy for your firm covering acceptable use, client data, disclosure and human review, then adapt it with your own counsel before you adopt it.

• Intake Pilot

  For US personal-injury firms: every inquiry gets your booking link inside a minute, three follow-ups over seven days, and a weekly report of captured, qualified and booked consultations.