ShiFt Answer Bank

Does AI intake count as legal advice?

A plain answer on the line between intake and the unauthorized practice of law.

Properly built AI intake does not give legal advice, and it must not. Capturing a caller’s name, situation, and contact details is intake. Telling a caller whether they have a case, what their claim is worth, or that the firm will represent them is legal advice — and from a nonlawyer tool that risks the unauthorized practice of law under Rule 5.5. A supervised AI redirects every advice question to an attorney with the same scripted line, logs it, and moves the caller toward a booking. The lawyer supervises the script under Rule 5.3.

  • •Gathering facts is intake; assessing the claim is legal advice — the tool must not cross that line.
  • •A nonlawyer giving legal advice risks unauthorized practice under Rule 5.5.
  • •The safe pattern is one scripted redirect, logged every time.

Where intake ends and advice begins

Intake is fact-gathering: who is calling, what happened, when, and how to reach them. That is administrative work a nonlawyer assistant has always been allowed to do. Advice is judgment about the law applied to those facts: whether there is a viable claim, what it might be worth, which strategy to pursue, or whether the firm will take the case. The moment a tool offers that judgment, it is doing something only a licensed lawyer may do. For an AI, the risk shows up in the seemingly helpful answer — “that sounds like a strong case” — which is exactly the sentence a supervised system is built never to say.

What a supervised tool says instead

When a caller asks “can I sue?”, “will you take my case?”, or “how much will I get?”, a supervised AI gives the same scripted redirect every time: an attorney will answer that, and here is how to get booked with one. It does not improvise, it does not estimate, and it does not commit the firm. Every redirect is logged, so the firm can show the pattern held. Under Rule 5.5 that keeps the tool clear of the unauthorized practice of law, and under Rule 5.3 the lawyer who approved the script stays responsible for the tool’s conduct. The script, not the caller, controls the conversation.

Confirm it before you go live

“No advice” is the second test in the Supervised Intake Standard, and the Bar-Safe AI Check scores whether your vendor’s tool actually redirects rather than answers. If you are not certain what your current tool says when a caller pushes for an opinion, that uncertainty is the finding — pull a few recordings, or send the vendor letter and get it in writing.

Questions answered

Full answers

Can AI intake tell a caller if they have a case?
No. Assessing whether a caller has a viable claim is legal advice, which a nonlawyer tool must not give. A supervised AI redirects that question to an attorney with a scripted, logged line and moves the caller toward booking.
Is AI intake the unauthorized practice of law?
Not when it is built correctly. Gathering facts is intake, which nonlawyers may do. It only risks unauthorized practice under Rule 5.5 if it gives legal advice, quotes fees, or forms a relationship — which a supervised tool is scripted never to do.

Related questions

AI answering and lead response

Attribution and revenue tracking

Qualification and speed to lead

Find your biggest revenue leak

The free leak check takes 2 minutes and shows you exactly where you're losing money.