What the opinion covers
It addresses generative AI under six obligations in the Model Rules: competence, confidentiality, communication, candor toward the tribunal, supervisory responsibilities and fees. ABA formal opinions are persuasive guidance rather than binding rules, and several state bars have issued their own.
What it means for AI intake
- Confidentiality: a prospective client's information is protected. Use a tool whose data handling you have read, whose terms do not train on your inputs, and get consent where the opinion says it is needed.
- Competence: know what the tool does well and where it fails. Do not let it improvise legal answers.
- Supervision: policies and training for staff, and a vendor agreement that holds the provider to your confidentiality duties.
- Communication: a prospective client should not be misled about who or what they are speaking with. Telling callers they have reached an automated assistant is good practice.
- Fees: if AI shortens the work, bill honestly for the time actually spent.
A checklist for choosing an intake AI
- It collects facts and books consultations. It does not give legal advice, and its script says so.
- Anything outside its scope goes to a person, with the whole conversation attached.
- Your data is not used to train the vendor's models, and you can see how it is stored and deleted.
- You have a written policy, staff training and a named person who reviews its work.
- Every conversation is logged so you can audit what was said.
- You have checked your own state bar's guidance, which may be stricter.
What we build
Our law firm system answers every potential client, qualifies and books the consult, hands anything sensitive to a person and keeps the record. The guardrails build adds test sets, redaction and audit logs for firms that want proof of how it behaves.